Climatology of Severe Winter Storms in Illinois

BULLETIN 53
STATE OF ILLINOIS
DEPARTMENT OF REGISTRATION AND EDUCATION
Climatology of Severe Winter Storms
in Illinois
by STANLEY A. CHANGNON, JR.
ILLINOIS STATE WATER SURVEY
URBANA
1969
BULLETIN 53
Climatology of Severe Winter Storms
in Illinois
by STANLEY A. CHANGNON, JR.
Title: Climatology of Severe Winter Storms in Illinois.
A b s t r a c t : Severe winter storms, those producing 6 inches or more snowfall and/or
damaging glaze, occur five times a year in Illinois, and cause more damage than other
forms of severe weather. The climatology of all severe winter storms in 1900-1960,
their damages, their synoptic weather conditions, and information applicable to design
and operational considerations are presented. Storms centered most often in northwestern
Illinois and averaged 6 inches of snow over 7500 square miles. Storm centers were
elongated with an axial ratio of 3:1 and WSW-ENE orientations. Point snowfalls of 12
inches and glaze with a radial thickness of 2 inches have occurred in all parts of Illinois.
Business losses resulting from winter storms are eight times the direct losses to property
and vegetation, and the greatest direct loss was $12.4 million in a 1924 storm. Five
synoptic weather types cause severe winter storms, and the Colorado low with a southerly
track produced 45 percent of all storms.
Reference: Changnon, Stanley A., Jr. Climatology of Severe Winter Storms in Illinois.
Illinois State Water Survey, Urbana, Bulletin 53, 1969.
Indexing Terms: applied climatology, blizzards, design storms, glaze, heavy snow, hydrologic budget, ice storms, Illinois, lake effect, severe storms, sleet, snow, storm damages,
storm models, synoptic climatology, wind with glaze.
STATE OF ILLINOIS
HON. RICHARD B. OGILVIE, Governor
DEPARTMENT OF REGISTRATION AND EDUCATION
WILLIAM H. ROBINSON, Director
BOARD OF NATURAL RESOURCES AND CONSERVATION
WILLIAM H. ROBINSON, Chairman
ROGER ADAMS, Ph.D., D.Sc., LL.D., Chemistry
ROBERT H. ANDERSON, B.S., Engineering
THOMAS PARK, Ph.D., Biology
CHARLES E. OLMSTED, Ph.D., Botany
LAURENCE L. SLOSS, Ph.D., Geology
WILLIAM L. EVERITT, E.E., Ph.D.,
University of Illinois
DELYTE W. MORRIS, Ph.D.,
President, Southern Illinois University
STATE WATER SURVEY DIVISION
WILLIAM C. ACKERMANN, Chief
URBANA
1969
Printed by authority of the State of Illinois – Ch. 127, IRS, Par. 58.29
C O N T E N T S
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data and analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Storm identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Page
. 1
. 1
.
2
. 3
. 3
.
3
. 5
P a r t 1 . Climatography of all storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Temporal characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Annual frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Monthly frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Daily frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Time of occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Storm point durations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Areal-geographical characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Areal distribution of storm cores . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motion of storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Areal extent of storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Characteristics of snowfall cores . . . . . . . . . . . . . . . . . . . . . . . . . . .
Damage information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Types of damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Storm of 26-27 January 1967 . . . . . . . . . . . . . . . . . . . . . . . . .
Areal-temporal distributions of damaging storms . . . . . . . . . . . . . . . . . . . . . .
Extremely damaging storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Meteorological conditions with severe winter storms . . . . . . . . . . . . . . . . . . . . . . . .
Storm classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Weather types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Detailed synoptic analysis of January 1967 storm . . . . . . . . . . . . . . . . . . . . . .
Frequency of weather types . . . . . . . . . . . . . . . . . . . . . . . . . .
Effect of Lake Michigan on severe winter storms . . . . . . . . . . . . . . . . . . . . . .
Patterns of selected winter storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6
6
6
7
7
7
8
8
8
10
10
12
13
13
15
15
16
17
17
19
20
20
21
22
Part 2. Climatography of glaze storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Temporal characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Annual frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Monthly frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Time of occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Storm point durations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Areal-geographical characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Geographical distribution of glaze areas . . . . . . . . . . . . . . . . . . . . . .
Areal extent of glaze areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Motion of storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Orientation of glaze areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Meteorological conditions with glaze storms . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
29
29
29
29
29
29
30
30
30
33
33
33
Part 3. Design information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Point data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Point frequencies of heavy snowfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Record high 2-day snowfall values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Maximum point snowfalls per storm and year . . . . . . . . . . . . . . . . . . . . . . .
Maximum snow depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Maximum glaze accumulations . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Glaze-wind relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Average frequencies of days with heavy snow, glaze, and sleet . . . . . . . . . . . .
Areal data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Area-depth relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Storm model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Regional probabilities of severe storm centers . . . . . . . . . . . . . . . . . . . . . .
Regional probabilities for successive storms . . . . . . . . . . . . . . . . . . . . . . . . .
Persistence of severe snow and glaze conditions in storm centers . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
34
34
34
35
35
36
36
38
38
41
41
42
42
43
43
45
Climatology of Severe Winter Storms in Illinois
by Stanley A. Changnon, Jr.
ABSTRACT
Severe winter storms, which usually produce snowfall in excess of 6 inches and often are
accompanied by damaging glaze, normally occur five times a year in Illinois, and they produce
more damage than any other form of severe weather including hail, tornadoes, or lightning.
In the 1900-1960 study period these storms were most frequent in January, but the dates
when storms were most frequent were 23, 24, and 25 December, with storms occurring on
these dates in 15 percent of the years.
The centers of the storms occurred most frequently in northwestern Illinois, and the
average storm produced at least 1 inch of snow over 32,200 square miles and more than 6
inches over 7500 square miles. The storms generally moved from the southwest, and the
storm centers were elongated with an axial ratio of 3:l and WSW-ENE orientations.
Damaging glaze was produced in 92 storms in the 61-year period. Glaze-producing
storms occurred most frequently in December, and the centers of glazing were usually in
central Illinois. Glazing with a thickness of 2 inches has occurred in most areas of Illinois.
Point snowfalls of 12 inches or more have occurred in all state areas, and a maximum of 37
inches at a point occurred in one 1900 storm. Many other results pertaining to all types of
design problems relating to severe winter storms are also presented.
The study of storm damages in the 1910-1960 period revealed that the most damaging
storm occurred on 17-19 December 1924 and produced $12,400,000 in property damages.
Unmeasurable business losses resulting from severe storms are estimated to be eight times
the direct losses to property and vegetation.
Five weather types lead to the production of the severe winter storms in Illinois. The
Colorado low with a southerly track produced 45 percent of the winter storms, and generally
produced the most widespread heavy snows, more damaging glaze conditions, and more
overall damage than did the other weather types.
INTRODUCTION
A detailed climatological study of all severe winter
storms occurring in Illinois during the 1900-1960 period
has been pursued to obtain extensive information concerning these frequently quite damaging snow and ice
storms. This study provides information that enlarges
our knowledge of the basic climatological aspects of
winter storms, statistics concerning the amount and types
of damage they produce, descriptions of the meteorological conditions producing these storms, and data helpful
in the design and planning for these events.
Severe winter storms in Illinois produce more total
damage than any other form of short-term severe weather
including hail, lightning, and tornadoes. Specific, complete dollar-loss figures are impossible to obtain for most
severe winter storms, and this makes an exact evaluation
of their damage-capability difficult. Furthermore, certain
forms of economic loss associated with winter storms,
such as reduced commercial operations resulting from
absenteeism caused by transportation stoppages, is extremely difficult to assess. Deaths and injuries to humans have resulted from many Illinois winter storms. A
large percentage of the property damages are experienced
by commercial carriers, communication media, utility
companies, and by various vegetation (figure 1).
It appears reasonable to conclude that, on the average,
severe winter storms each year produce more than $2
million in property damage, 4 deaths, and at least 40
injuries, plus an untold amount of loss to commercial
operations and extensive personal hardships. In comparison, the annual average losses for tornadoes are $1.2
million and 5 deaths 1 ; those from lightning are $0.1 million and 6 deaths 2 ; and average hail losses amount to
$2.2 million and no deaths.3 Severe winter storms do contribute sizable amounts of moisture for storage in surface
reservoirs, and the snow cover protects winter crops and
helps prevent erosion.
1
Figure 1. Results of severe winter snowstorms in Illinois (Chicago Tribune photo)
This report is organized into three major parts, preceded by a section describing the data employed to define the 304 storms in the 61-year period and the types
of analyses used. Part 1 of the report is primarily a climatography of the 304 winter storms. In this, their temporal variations, including hourly, daily, monthly, and
annual frequencies, are described along with their geographical patterns and areal frequencies across Illinois.
One section is devoted to statistics on storm damages including the worst storms, types of damages, and weather
conditions related to the most damaging storms. Another
section includes a detailed study of the meteorological
conditions that produced these storms, and a final portion
concerns snow and glaze patterns exhibited by various
unusual severe winter storms.
Part 2 presents a comparable study of the 92 storms
that were classified as glaze storms. Most of these produced widespread damaging glaze, either with or without
large snowfalls. A separate climatography of glaze storms
was made because many produced quite extensive damages. The climatological statistics presented for these
storms include various temporal and areal frequencies,
damage data, and information on the meteorological con2
ditions that produced them.
Severe storm information useful in designing structures
and in operational planning is the subject of Part 3 of
the report. Certain data presented in Parts 1 and 2 are
summarized in this final section. Averages and extremes of
storm frequency, snowfall depth, and ice thickness are
presented for points throughout Illinois. Models of severe
winter storms are shown, and regional statistics are offered concerning the probabilities of a sequential series
of storms, persistence of post-storm damaging conditions,
and areal and temporal likelihoods of storms.
Data from a detailed analysis of a very severe winter
storm in January 1967 were incorporated in various sections of the report to provide certain comprehensive information that was not available for the storms in the
1900-1960 period.
Acknowledgments
This report was prepared under the direct supervision
of Glenn E. Stout, Head of the Atmospheric Sciences
Section, and under the general supervision of William C.
Ackermann, Chief of the Illinois State Water Survey.
Special credit is due Harold Q. Danford who performed
many of the synoptic weather analyses for the 304 winter
storms. During the five years that this research has been
in progress, several persons were involved in the difficult
and sizable task of data reduction and map plotting.
Among those who greatly assisted in these activities were
I-min Chiang, David Pingry, Ruth Braham, and Edna
Anderson. Helpful suggestions on the contents of this
report have come from Floyd A. Huff, Hydrometeorologist on the Survey staff, and from David M. Hershfield
of the Agricultural Research Service. Since this report
could not have been compiled without availability of the
original records of the U. S. Weather Bureau stations in
Illinois, the assistance of William L. Denmark, State
Climatologist for Illinois, in making these data readily
accessible is gratefully acknowledged. The entry of the
storm data in punch cards and the ensuing machine
analysis were performed under the supervision of Marvin
C. Clevenger, Machine Supervisor on the Survey staff.
Mrs. J. Loreena Ivens, Technical Editor, edited the report. John W. Brother, Jr., Chief Draftsman, and William
Motherway, Jr., Draftsman, prepared the illustrations.
Special thanks are due the many businesses and commercial organizations and their personnel who supplied detailed damage data for the January 1967 storm.
DATA AND ANALYSES
Data
A major source of data for this study was the climatological data publications of the U. S. Weather Bureau
issued monthly in the 1900-1960 period.4 In many of
these issues weather summary texts appear, and these
contained descriptions of many damaging winter storms,
including the amount and types of damage. Old newspapers also were studied to obtain additional data on
the amount of damages. Data on the occurrence of 6-inch
or greater snowfalls in 48 hours and on daily glaze occurrences were obtained from an examination of the
published monthly issues of Climatological Data, from a
computer analysis of the daily snowfall data entered on
IBM cards for 62 Illinois stations with long records,5 and
finally from a search of the original, unpublished weather
records of all other U. S. Weather Bureau stations in
Illinois.
Additional weather data relating to each storm came
from these unpublished records of the Weather Bureau
substations. In addition to the amounts of snowfall and
the glaze occurrences, data obtained included the beginning and ending times of the snow or glaze, local
damages, persistence of damaging snow and glaze conditions after the storm, and the occurrence of other
weather phenomena such as sleet and high winds. Considerable data on extreme glaze conditions were found in
research reports dealing with glaze.6, 7
Storm Identification
Before the analyses of severe winter storms could be
initiated, quantitative criteria to identify their occurrence
were needed. The most desirable identification would
have been that based on the amount of damages produced by a winter storm. Unfortunately, such data were
not available in the weather records for many obviously
damaging storms, so definitions based on available,
quantitative weather data (such as the amount of snow
or extent of glaze) were devised.
The snowfall amounts and areal frequencies of glaze
and sleet reports associated with several storms that had
known large losses were identified and analyzed. This
analysis revealed that in most instances damaging
(severe) winter storms were those that produced 6
inches or more snowfall in 48 hours or less at some point
in Illinois and/or those in which glaze covered at least
10 percent of the state. This agrees with the U. S.
Weather Bureau’s heavy snow warning which is issued
for an area when a fall of 6 inches or more is expected
in a 24-hour period. Sleet often accompanied glaze and
snow storms but was not a separable damaging event.
The historical storm analyses also showed that all damaging snow or glaze storms began and ended within Illinois during a 60-hour period or less, and were the result of a single, definable set of meteorological conditions.
In cases when heavy snow and glaze continued intermittently over 4-, 5-, or 6-day periods, there was actually
a series of two or three storms, each produced by a different set of meteorological conditions.
The severe winter storms in Illinois were identified
and selected by one or more of these established criteria: 1) a snowstorm that produced 6 inches or more
snowfall at a point in 48 hours or less; 2) a snowstorm
that produced conditions leading to property damages,
deaths, or injuries regardless of the amount of snowfall;
3) a glaze storm in which 10 percent of the cooperative
U. S. Weather Bureau substations in Illinois reported
glaze; and 4) a glaze storm in which property damages,
deaths, or injuries occurred. The second definition had
to be included because a few storms in which the maximum point snowfall in Illinois was only 4 or 5 inches
produced damages because the snows were associated
with high winds resulting in drifting sufficient to cause
traffic stoppages and auto accidents.
3
a. Weather data available
for 8-9 January 1920
b. Patterns of snowfall,
glaze, and sleet
c. Times of beginning and ending
of storm and forward advance
Figure 2. Data and analysis of severe winter storm on 8-9 January 1920
4
Analyses
The available data for each storm were plotted on an
Illinois base map, as illustrated in figure 2a. A text also
was prepared for each storm, describing all other available information, including the amount and type of damage. Snow isolines and areas of glaze were circumscribed
( figure 2b ).
Several desired storm statistics then were derived from
these maps and texts. Storm beginning times at all points
were used to calculate the storm motion across the state
(figure 2c). The storm-start isochrones for the storm on
8-9 January 1920 reveal that it moved from the south.
The time of each storm start within Illinois also was determined from the storm maps.
The areas of 6-inch or greater snowfalls were labeled
as storm ‘cores’ as a means of objectively identifying
areas of storm maximization. Several measurements of
these cores were made, including their length, which was
defined as the greatest straight-line distance within the
area; their orientation, which was that of the length
measurement; and their width, which was an average
determined from three width measurements taken at right
angles to the length at different locations. In core number
1 (figure 2b) the length was 92 miles, its orientation
was 225 degrees, and the average width based on the
three indicated width measurements was 15 miles. In
certain very extensive storms, or those near the state
boundaries, the length and width of cores could not be
determined because the 6-inch isolines extended beyond
Illinois (core number 2 in figure 2b). This limitation affects the mean calculated for the lengths and widths of
cores, making them slight underestimates. As shown in
figure 2b, many storms had two or more cores, and
measurements were made for each.
A storm duration was the average at a point as determined from the durations at the stations in the center
of the storm (in and around the heaviest snowfall
amounts, or if little snow, at the center of the glaze
area). In the example (figure 2c), the beginning and
ending times were available at only three stations, and
the storm duration was 13 hours.
The persistence of deep snow cover and/or damaging
glaze at each station in the center (core or cores) of the
storm was listed. These values were used to calculate an
average persistence value for each storm, which for the
8-9 January 1920 storm was 2 days. Also listed for each
storm was the number of stations reporting glaze, sleet,
and 6-inch snowfalls.
To ascertain whether a given storm qualified as a glaze
storm, regardless of the amount of snow, there had to be
recorded damages from glaze or the number of glaze reports had to equal or exceed 10 percent of the number of
reporting stations that year in Illinois. The number of reporting stations was 70 in the 1900-1920 period, 90 in the
1921-1940 period, 130 in the 1941-1950 period, and 170
in the 1951-1960 period. Thus, a 1910 storm qualified as
a glaze storm if it had 7 stations with glaze, but a storm
in 1955 had to have 17 stations with glaze to qualify as a
glaze storm, unless a record of damages due to glaze was
found for the storm. The illustrated storm (figure 2c)
did not have sufficient reports of glaze to qualify, but
there was reported damage from glaze, thus qualifying
it as a glaze storm as well as a snow storm.
The damage data available for each storm also were
summarized. The status of glaze damages was classified
as an occurrence, as a nonoccurrence, or as unknown.
The total amount of storm damage (resulting from snow
and glaze) was classed into one of four categories. One
class was labeled ‘no damage-unknown damage’ and consisted of storms with less than $1000 or those for which
no damage data could be found. The ‘minor’ damage
class was assigned to storms producing $1000 to $10,000
in damages; the ‘moderate’ class was for those producing
$10,000 to $200,000; and the ‘extreme’ class included
those with damages greater than $200,000 and/or those
producing one or more deaths to humans. The illustrated
storm (figure 2b) was classed in the moderate group. To
make the damages for all storms comparable, the total
dollar losses of each storm were normalized to the value
of the 1960 dollar. This normalization was based on annual price indices for building materials.8
The study of the synoptic weather conditions related
to the occurrence of each of the 304 storms was based
largely on analysis of the published daily surface weather
maps of the U. S. Weather Bureau. Since a uniform, comparable classification system was desired for all storms,
an upper air analysis could not be incorporated because
of the lack of such data for much of the 1900-1940 period.
Further details about the analytical approach employed
are presented in the sections describing these results.
Certain synoptic weather data relating to a winter storm
in 1967 were employed to reveal detailed information not
available for all storms.
5
Part 1. Climatography of All Storms
A snow or glaze storm could be defined as a severe
winter storm on the basis of four possible conditions occurring in 60 hours or less. Many storms qualified by
fulfilling more than one of these four criteria, and thus
many combinations of the four criteria could and did
exist in certain storms. The criteria associated with the
304 storms formed 10 combinations (table 1). The greatest number in a combination is 125 storms, and these
qualified because they each produced 6 inches or more
snowfall, but no known snowfall damages, no glaze damages, and no widespread glaze.
The other three classes with relatively large total numbers of storms were the 6-inch snow and damage (86
storms), the class with all four conditions (32 storms),
and the class with known damages, widespread glaze,
and glaze damage (24 storms). The 24 storms in this last
class were generally rather extensive damaging glaze
storms without snow, whereas those in the snow-damage
class were simply snowstorms without any glaze. The 32
storms with all four conditions were truly ‘severe’ storms,
fulfilling all definitions of a severe winter storm.
Examination of the frequency of the different storms in
each month reveals that a relatively large percentage of
the December storms fell into the snow-without-knowndamage class and in the glaze-damage class. Relatively
few of the February and March storms were in the class
defined by a 6-inch snowfall without damage or glaze.
Combining all storms in which the 6-inch snowfall
criterion was met reveals that in 269 of the 304 storms at
least one point in Illinois had that amount of snow. A
total of 138 of these 269 storms had known damages, 40
were associated with widespread glaze conditions, and
53 also had glaze damage.
Slightly more than 50 percent of all the storms, 173,
were identified as having known damages from snow
and/or glaze conditions. Thirty-five of these 173 knowndamage storms were associated with storms having point
snowfalls of less than 6 inches. Eighty-seven of the 304
TEMPORAL
Number of storms
Oct Nov Dec Jan Feb Mar Apr
Criteria
combinations
May Total
≥6
inches of snow
(damages unknown
and no glaze)
≥ 6 inches snow and
known damages
(no glaze)
≥ 6 inches snow,
known damages, and
widespread glaze (no
glaze damage)
≥ 6 inches snow,
known damages,
widespread glaze,
and glaze damages
≥ 6 inches snow and
widespread glaze
(no damage)
≥ 6 inches snow,
widespread glaze,
and glaze damage
(amount of damage
unknown)
≥ 6 inches snow,
known damages,
glaze damage (glaze
not widespread)
Known damages,
widespread glaze,
glaze damage (snow
< 6 inches)
Known damages,
glaze damage (glaze
not widespread,
snow < 6 inches)
Known damages (no
glaze and snow
< 6 inches)
1
7
30
34
26
19
7
1
125
0
7
13
18
20
25
3
0
86
0
0
0
0
0
1
0
0
1
0
1
6
13
8
3
1
0
32
0
0
0
2
0
2
0
0
4
0
0
1
1
0
0
0
0
2
0
1
5
1
6
4
2
0
19
0
0
9
8
3
3
1
0
24
0
1
1
2
1
5
0
0
10
0
1
0
0
0
0
0
1
0
storms had glaze damages, but only 58 of the 87 with
damaging glaze were associated with widespread glaze
conditions. Widespread glaze conditions occurred in 63
storms, and 39 of these storms also produced point snowfalls of 6 inches or more.
CHARACTERISTICS
Annual Frequencies
The average number of severe winter storms in Illinois
during a year is 5, but there have been as many as 12
storms in one year (1912) and as few as 1 storm in one
year (1919 and 1925). Maximum and minimum frequencies of storms for periods ranging from 1 to 10
years appear in table 2. In an average 10-year period, 50
severe winter storms would occur, but the 10-year extremes show 68 occurred in the 1942-1951 period (18
6
Table 1. Number of Severe Winter Storms in Various
Classifications Based on Combinations of Definitions
Table 2. Maximum and Minimum Number of Storms
for Various Year-Periods
Number of storms for various periods
3-year
2-year
5-year
Storm
occurrences
1-year
Maximum
(year)
12
(1912)
25
68
35
19
(1950-51) (1943-45) (1943-47) (1942-51)
Minimum
(year)
1
(1919,
1925)
16
36
8
5
(1902-03, (1919-21) (1919-23) (1916-25)
1907-08,
1919-20,
1937-38)
10-year
as 5 storms have occurred in March (1912), and 4 storms
have occurred in a single December (1915 and 1951)
and in January (1918 and 1957). There have been cases
with no severe winter storms in all the cold season
months. Probabilities for the occurrence of 1, 2, 3, or
more severe winter storms in each month are presented
in table 4. There is a 70 percent chance for 1 or more
severe storms to occur somewhere in Illinois during
January and a 43 percent probability for 2 or more
storms in any January. February and March both have
probabilities of about 60 percent for 1 or more storms.
Figure 3. Annual number of severe winter storms
above average) and 36 occurred in 1916-1925 (14 below
average).
The annual number of storms in 1900-1960 are plotted
in figure 3. This time distribution shows that 3-year or
longer low frequency periods occurred in 1905-1908,
1919-1923, and 1937-1941. Three-year or longer high
frequency periods occurred in 1909-1912, 1934-1936, and
1943-1945. The annual frequency data were tested for the
existence of cycles in their occurrences, but no evidence
of a cyclic distribution was found.
Monthly Frequencies
The total number of storms in each month is shown in
table 3. The earliest storm occurred on 28-30 October
1925, and this was the only severe winter storm to occur
in October. The latest storm in spring occurred on 1-2
May 1929, and this was the only severe winter storm to
occur in May. January, with 79 storms, had the greatest
number, and numbers in December, February, and
March were almost equal. The numbers in November
and April were of comparable magnitude. The four
months beginning with December have yearly averages
of 1 storm, although the January value is somewhat
greater than 1 and in an average 10-year period 13 winter storms occur in January.
The monthly extremes (table 3) reveal that as many
Table 3. Average and Extreme Monthly Number of Storms
Sorm
ocurrences
Oct
Nov
Total, 1900-1960
1
18
Average per year
Maximum in 1 year
Minimum in 1 year
*Considerably less than 1
*
1
0
*
2
0
Number per month
Dec Jan Feb
Mar
65
1
4
0
79
1+
4
0
64
1
3
0
62
1
5
0
Apr
14
*
2
0
May
1
*
1
0
Table 4. Probability for Severe Winter Storms
Number of severe
storm ocurrences
Nov
Dec
1 or more
2 or more
3 or more
25
5
<1
67
25
12
Probability (percent)
Mar
Jan
Feb
70
43
15
62
31
12
60
30
7
Apr
Year
21
2
<1
100
96
87
Daily Frequencies
The number of times that a severe storm occurred on
each date in the 15 November to 19 April period was
determined and then plotted (figure 4). The distribution
of the occurrences is not uniform with certain periods
having relatively higher frequencies than dates on either
side of these periods (27-28 November, 22-26 December,
25-26 February, 2-3 March, 10-12 March, 18-20 March,
and 25-26 March). The daily frequencies in January
and most of February vary but do not exhibit any marked
high incidence periods. Notable periods of low incidence of storms include 3-4 December, 15-17 December,
3-5 January, 23-28 January (time of the well-known January thaw), 20-24 February, 15-17 March, and 21-24
March. The high frequencies on 24, 25, and 26 December and on 2-3 March equate to probabilities of 15
percent for a severe storm occurrence on each of these
dates.
Time of Occurrence
The time that each storm initiated in Illinois could be
identified for 284 storms. The number of initiations in
each hour was determined, and the hourly distributions
appear in figure 5. The preferred 4-hour period of initiations was from 0900 to 1300 CST when 33 percent of the
Figure 4. Number of times a severe winter storm occurred on each date somewhere in Illinois
7
Figure 5. Time of initiations of severe winter storms in Illinois
storms began; in general, most storms began in the daylight hours.
Examination of the hourly distributions of initiations
in each month revealed no significant differences from
that shown in figure 5 for all the storms. The peak shown
for the hour ending at 0100 was a result of initiations at
this time occurring in December, January, and February.
The diurnal distribution of hours when measurable snowfall occurs in most of Illinois also shows a peak of activity
in the daylight hours.9
Storm Point Durations
The average point duration could be determined for
245 of the 304 storms. As shown in table 5, these values
ranged from a low of 2.0 to a high of 48.0 hours, with a
median of 14.2 hours. The average duration of all measurable snowfalls in Illinois varies from 5.1 hours in
southern Illinois to 6.3 hours in northern Illinois.9
More than 80 percent of the duration values in each
month were less than 20 hours, but the distribution of
Figure 6. Frequency distribution of average point durations of
severe winter storms
Table 5. Median and Extreme Point Durations
for 245 Severe Winter Storms
Duration (hours)
Storm
duration
Maximum
Minimum
Median
Nov
Dec
Jan
Feb
Mar
Apr
All
storms
35.0
3.0
9.0
48.0
2.0
14.4
35.9
2.5
14.0
47.0
3.0
12.3
42.0
2.0
13.1
34.0
4.7
8.9
48.0
2.0
14.2
values was quite skewed, with a few storms in the 30hour or longer range in each month. Therefore, mean
monthly durations were less meaningful than were
median durations. The skewness of the duration distribution is revealed in figure 6, which is a histogram based
on all 245 storms. Sixty percent of all storms had average
point durations between 4 and 16 hours, whereas 10 percent had durations ranging from 24 to 48 hours.
The median monthly durations (table 5) reveal that
storms in December and January were somewhat longer
lived at a point than those in other months. Durations of
storms in November and April were notably less than
those in the other months.
AREAL-GEOGRAPHICAL CHARACTERISTICS
Areal Distribution of Storm Cores
The occurrences of 6-inch snowfall cores and damaging
glaze areas in each of the nine crop reporting districts in
Illinois were tabulated on a monthly and total basis. The
total based on the number in each district were considerably greater than the 304 storm total because 1)
8
many storms had two or more snowfall cores (see table
8) as well as glaze areas, and 2) a core often extended
over three or four districts and was counted as an occurrence in each district.
The monthly patterns in figure 7 exhibit latitudinal
distributions. The NE district has a maximum frequency
in November, December, and April. Both southern dis-
a. November
b. December
c. January
d. February
e. March
f. April
Figure 7. Monthly number of times a 6-inch snow core or glaze damage urea occurred in 1900-1960 period
9
(Cook and Du Page Counties) ; the third is in southcentral Illinois; and the fourth is in extreme southeastern
Illinois. Quite low incidence areas occur in central Illinois (Sangamon County) and in east-southeastern
Illinois.
The number of times an area of 6-inch or greater snowfall occurred in each grid rectangle (figure 9a) was
counted for all 269 storms with 6 inches or more snowfall. The pattern based on these values (figure 9c)
reveals a general latitudinal distribution. However, the
area of 40 or more occurrences extends into west-central
Illinois and two separate highs appear in northern Illinois.
The area of least frequency is in extreme southern Illinois, although a minor high defined by the 25-isoline
appears in southeastern Illinois.
Motion of Storms
Figure 8. Number of times a 6-inch snow core or glaze damage
area occurred in 1900-1960 period
tricts have low occurrences in November and April; in all
other months, the SE district has the lowest number of
storm core occurrences. The NW district led all districts
in number of cores in January, February, and March.
The March pattern reveals less latitudinal distribution
than in other months with relatively high frequencies of
storm cores in the W and WSW districts. Although the
areal distribution is generally latitudinal in January,
February, and March, there is also a tendency for a decrease from west to east.
The total number of core (snow or glaze) occurrences
in each district during the 1900-1960 period is shown in
figure 8. This pattern also reflects the latitudinal northsouth decrease through Illinois, as well as the west-east
decrease across Illinois.
The location of the maximum snowfall value in each
storm producing 6 inches or more (total of 269 storms)
was identified and plotted. On this map, a grid consisting
of rectangles with a width of 30 seconds in longitude
and a length of 30 seconds in latitude (about 935 square
miles) was constructed (figure 9a), and the number of
storm maximums in each rectangle was calculated. The
pattern based on these grid values appears in figure 9b,
and four areas where storm maximum snowfalls have
frequently occurred are indicated. The major high in
Illinois is oriented SW-NE in northwestern Illinois. The
second high occurrence area is in northeastern Illinois
10
The motion of the storms across Illinois could be ascertained for 283 of the 304 storms. These motions were
labeled as one of 360 possible degrees of the compass
and were sorted and counted for 10-degree intervals
(figure 10). A large number moved from the SW and
WSW, with 47 percent moving from azimuths in the
216-255 degree sector. Minor maximums appear for
storms with southeasterly motions (136-165 degrees) and
for those with northwesterly motion (306-335 degrees).
The three motion maximums shown for all storms (SW,
SSE, and NW) also appear in motions found in all months
except November, which has only a maximum of motions
from the southwest. The one 10-degree sector from which
storms moved most frequently in all months is 236-245
degrees.
Areal Extent of Storms
The areal extent of snowfall of 1 inch or more was
determined for the 269 storms with point amounts of 6
inches or more. Average values for different snowfall
increments and each month are presented in table 6. The
totals shown for each month reveal that, on the average,
the April snowstorms are about 30 percent smaller than
those in the other five months. The average December
Table 6. Average Areal Extent of Snowfall in 269
Storms with 6 Inches or More Snow
Average number of square miles
All
storms
Snowfall
(inches)
Nov
l–2
2–4
4–6
6–8
8–10
1 0 – 12
> 12
7,540 6,200 6,100
5,740 6,260
9,915 11,920 10,820 10,660 10,820
7,100 8,720
6,970 7,750 7,700
3,530 4,490
4,005 4,490 4,170
2,030 1,990
2,190 1,975 2,295
710
565
1,170
925
990
215
700 1,205
545
355
6,240 6,510
6,785 10,380
4,280 7,530
2,940 4,160
1,370 2,070
945
915
140
640
29,380 34,160 33,395 33,205 32,620
22,700 32,205
Total
Dec
Jan
Feb
Mar
Apr
a. Grid overlay based on
30-second by 30-second
divisions of latitude
and longitude
b. Distribution of maximum
storm snowfall values
in 269 snowstorms with
6 inches or more
c. Distribution of 6-inch
or greater snowfall areas
Figure 9. Distribution of maximum snowfall, 1900-1960, based on number per grid square
snowstorm has the largest monthly value, but it is not
significantly different from those in January, February,
and March. However, the average areal extents of heavy
snow, 6 inches or more, in January, February, and March
storms are almost 1000 square miles greater than that in
December.
The total value based on all storms reveals that, in the
average severe snowstorm, 32,205 square miles are cov-
Figure 10. Direction of storm movement across Illinois sorted
by number per 10-degree interval
ered by 1 inch or more snow, which is 57 percent of
Illinois. Furthermore, 7785 square miles experience 6
inches or more snowfall and 1555 square miles have 10
inches or more. In December, January, February, and
March there has been at least one snowstorm that covered the state (56,400 square miles) with 1 or more
inches of snow. The most extensive snowstorm in November produced 1 inch or more snow over 50,900 square
miles, and the most extensive April storm spread over
54,700 square miles.
The areal extent of glaze areas was more difficult to
assess than that for snowfall because of the lack of dense
point observations needed to define this either yes-or-no
event. To estimate the areal extent of glaze, the number
of Weather Bureau stations reporting glaze was counted
for each storm. Because of the generally uniform distribution of these stations, each station was considered to
represent a certain percent of the state. Since the number
of stations capable of reporting glaze increased during
the 1900-1960 period, the percent of state area assigned
to a station was varied according to the number of stations. A station glaze report in the 1900-1920 period represented 2 percent of the state; 1 station in the 1921-1940
period represented 1 percent; 1 in the 1941-1950 period
was 0.8 percent; and a report in the 1951-1960 period
represented 0.5 percent of Illinois. Thus, the extent of
11
glaze in a storm in 1912 with 5 stations reporting glaze
was estimated as 10 percent of Illinois, whereas the extent in a 1955 storm with 5 stations reporting was 2.5
percent.
These values were sorted into percentage intervals by
month, and the monthly and total distributions are
shown in table 7. No glaze occurred in nearly half ( 141)
Table 7. Percent of Illinois Covered by Glaze
in 304 Severe Winter Storms
Percent of
Illinois
OctNov
0
14
1–2
3
3–5
0
1
6–9
10–15
1
16–20
0
0
21–25
26–30
0
31–40
0
41–50
0
Maximum
single-storm
p e r c e n t a g e 11
Number of storms
Dec
Jan
Feb
Mar
AprMay
Total
8
4
1
0
1
0
1
0
0
0
141
63
20
16
31
9
6
5
9
4
25
28
11
5
5
7
1
1
2
3
2
31
18
4
2
10
4
1
2
5
2
32
14
4
3
6
2
2
0
1
0
28
13
6
5
6
2
1
1
0
0
46
43
3 2
28
of the 304 winter storms, and in 83 other storms the glaze
area covered less than 6 percent (3300 square miles) of
the state. In a few storms, the glaze area was quite extensive, reaching a maximum estimated state coverage of
46 percent for a single storm during 13-14 December
1937 (see figure 26).
of multicore storms in each month represented a percentage of the total monthly storms with cores that was
relatively the same in all months. For instance, the number of storms with two or three cores was 29 percent of
October-November storms with cores, 30 percent of
December storms, 36 percent of January storms, 33 percent of February storms, 44 percent of March storms, and
21 percent of April-May storms. Only March with relatively more multicore storms and April-May with relatively few, showed percentages somewhat different from
those in the other months.
The average widths of 65 of the 387 cores could not
be determined because either very little of these cores
was in Illinois or one side was beyond the Illinois boundary. Although both ends of some cores were not in the
state, width values were calculated for these if the portion of the core in Illinois was at least 60 miles long.
Hence, more measurements could be made for core
widths than for lengths; the lengths of 274 of the 387
cores could be measured.
The number of width and length values in various
class intervals of miles appears in figure 11. Nearly 38
Characteristics of Snowfall Cores
The snowfall cores, defined as areas with 6 inches or
more snowfall, were analyzed for the number per storm
and their size, shape, and orientation. Many storms had
two or three cores, and in some storms the dimensions of
some large cores could not be measured accurately, since
the core boundaries extended beyond Illinois.
The number of storms with varying numbers of snowfall cores is shown in table 8. Ninety-three of the 269
storms with cores had two or three cores, but 176 storms,
65 percent of the total, had only one core. The number
Table 8. Number of 6-Inch or Greater
Snow Cores per Storm
Number of storms with
various numbers of cores
No core
1 core
2 cores
3 cores
Oct-Nov
Dec
Jan
Feb
Mar
Apr-May
2
10
10
4
8
1
12
39
44
40
30
11
4
10
20
14
17
3
Total
35
176
68
1
6
5
6
7
0
25
Month
12
Total
cores
Total
storms
with
cores
23
77
99
86
85
17
17
55
69
60
54
14
387
269
Figure 11. Distributions of lengths and widths of storm cores
percent of the 274 cores had lengths in the range from
1 to 50 miles, and 67 percent had lengths of less than 104
miles. The median length value was 68 miles, the longest
was 325 miles, and the shortest was 10 miles. It should
be realized that the median length value represents that
for cores defined entirely in Illinois and many bigger
cores extending across Illinois were not measured, making
this median an underestimate for big cores.
The mode of the width values was the 11- to 20-mile
class (figure 11). The greatest width value measured
was 233 miles, the smallest was 5 miles, and the median
value was 22 miles.
The frequency distributions of widths and lengths in
each month did not differ greatly from the total distributions shown in figure 11. However, the cores in November, March, and April storms generally were narrower
than those in the other months.
The orientation of the core (major axis) was also determined for all 387 cores in the 269 storms. These were
sorted into 10-degree intervals ranging from 176-185
degrees (oriented N-S) through 346-355 (oriented NNWSSE). The number in each 10-degree interval was used
to construct the frequency diagram in figure 12. This
shows that the preferred orientation of the snowfall cores
was 246-255 degrees (WSW) where 72 cores occurred
(19 percent). Most of the cores, 63 percent, were oriented
in the 40-degree sector from 226 through 265 degrees.
Similar distributions were found for the orientations of
cores in all months.
Figure 12. Frequency of orientations of storm cores
DAMAGE I N F O R M A T I O N
A search of the storm damage records compiled for this
study revealed that the period of most complete records
was 1946-1955. In this 10-year period physical damages
from 16 storms in Illinois totaled $17 million with 30
known dead and at least 350 persons injured. However,
these 16 storms represented only about one-third of the
54 severe winter storms that occurred in the 1946-1955
period, and no reasonably accurate loss data were available for the other 38 storms.
Types of Damage
Severe winter glaze storms produce damages to communication and utility companies (largely through wire
and pole losses, figure 13); to vegetation (breakage of
branches and limbs); to building structures (broken
antennas and roof damage from ice loading and falling
tree limbs); and to vehicles (slick road surfaces resulting
in accidents). Severe winter snowstorms also produce
damages to vehicles (slick surfaces and poor visibility
both producing accidents) and to buildings (roof damages from weight of excessive snow).
As examples, a glaze storm on 27 November 1955 produced 1000 traffic accidents in Chicago, and in a storm
in northern Illinois on 13 February 1950, 64,000 houses
lost electricity and many had broken TV antennas and
were damaged from falling tree branches and excessive
ice plus snow loading on their roofs. Flooding has also
been a damaging by-product of a few winter storms
when either heavy snowfalls melted rapidly or when
heavy rain fell on heavily glazed surfaces, resulting in
rapid runoff. Although the winter storms studied are
those that occurred within 60 hours or less, many achieve
added losses, as compared with warm season forms of
severe weather, because the damaging snow or ice conditions persist at a point over much longer periods of
time than do all other forms of severe weather.
Both types of conditions can result in injuries and
losses of life in automobile accidents, falls on slick surfaces, freezing to death while isolated or trapped by
blizzard conditions, overexertion from snow removal,
and electrocution from fallen power lines. The Weather
Bureau estimates that 85 percent of ice (glaze) storm
deaths are traffic-related.
Either glaze or snow conditions often cause stoppages
or serious delays in all transportation modes. Slick roads
(glaze or snow), reduced visibility (blowing snow), and
blocked roads and railroads (drifting heavy snow) all
affect personal travel, business activities, and deliveries
of goods and products via automobiles, trucks, railroads,
and commercial air carriers. A 1951 study of traffic reduction on a main highway due to ice and snow conditions revealed that 40-percent reductions occurred on
Route 66 near St. Louis.7 Business losses resulting from
such delays may be partially recovered, but are hard to
account for. Therefore, such ‘indirect loss’ figures were
just not available. Such information was collected for a
recent storm in January 1967, and a description of this
storm is presented to illustrate the direct as well as the
indirect losses.
13
Figure 13. Examples of heavy glaze and damages to power lines, 26-27 January 1967 (courtesy
14
of
Central Illinois cooperative)
Storm of 26-27 January 1967
The most comprehensive records of direct damages,
deaths, and indirect losses ever assembled for a winter
storm were collected for this storm. This storm began in
Illinois at 0000 CST on 26 January and ended 26 hours
later. Snowfall in northeastern Illinois exceeded 23
inches (a new record for Chicago and environs), and
the associated glaze storm in central and eastern Illinois
(figure 13) was a record (radial thickness of ice on wires
was 1.7 inches in places). Snowfall exceeded 6 inches
over the northern half of the state (26,650 square miles),
and high winds produced blizzard conditions and excessive drifting. In many respects this is a design winter
storm for the northern half of Illinois in that it combined
excessive snow and glaze.
The losses resulting from various physical property
damages were sorted by location (table 9), and these
totaled nearly $22 million. However, estimates of the
indirect losses to business amounted to $174 million,
eight times the amount from direct losses. These data
furnish an unusual perspective of catastrophic loss with
$195 million in losses, 56 dead, and untold injured. Since
much of the damage occurred in Chicago, some of the
results of the storm in Chicago were summarized to
illustrate certain conditions attendant to many very
severe winter storms.
Table 9. Damages and Dollar Losses Resulting
from 26-27 January 1967 Storm in Illinois
storm was to find places to deposit the snow removed
from streets and sidewalks. Considerable snow was
loaded into empty freight trains and taken from the city
to be dumped in rural areas.
Many Chicago deaths were attributed to heart attacks
which occurred to persons removing the heavy snow
from passageways or attempting to dig out their stranded
automobiles. One small girl was shot and killed when
police attempted to stop a band of looters. All told, 31
deaths directly attributable to the storm occurred in the
city. Four were injured when two elevated trains collided during the storm, but an estimate of total injuries
was not readily available.
For the first time in history, all the Chicago schools
were closed. Though some homes were without heat,
most in the city had lights and telephones. In fact, while
all transportation was halted, 12 babies were delivered
using telephoned instructions. About 20,000 stranded cars
blocked expressways; no taxis could operate; over 3300
buses were unable to move; and even the slowly moving
subways were 40 percent incapacitated. Only rescue helicopters could use the three closed municipal airports.
The closing of almost all stores and factories was accompanied by minimal mail delivery, and newspaper delivery existed only in the Loop. Looting of stranded cars
and closed shops reached new heights, with one looting
per minute, as estimated by police. Governor Otto Kerner delayed all Chicago monetary transactions by declaring 28 January to be a legal bank holiday.
Direct losses from various physical damages
Location
Amount (dollars)
Chicago
Other downstate
urban areas
Rural power utilities
Central Illinois
communication companies
State total
3,500,000
5,661,000
11,029,139
1,623,000
21,813,139
Indirect losses to business
Location
Amount (dollars)
Chicago
Downstate
State total
150,000,000
24,000,000
174,000,000
Deaths
Location
Chicago
Downstate
State total
Number
31
25
56
The Chicago Transit Authority conservatively estimated their damages to be $1 million, and other commercial and private structures received physical damages
estimated at $2.5 million. Losses to Chicago businesses
exceeded the $150 million mark. The winds during the
storm averaged 25 mph with gusts to 70 mph causing
10-foot high drifts from the 23-inch snowfall and also
limiting visibility to zero. A serious problem after the
Areal-Temporal Distributions of Damaging Storms
A total of 131 storms was classed in the category of no
damage-unknown damage, and because of the uncertainty of the damage data for many storms in this group,
parameters of these storms were not analyzed. The minor
damage category ($1000 to $10,000) consisted of 81
storms; the moderate damage category ($10,000 to $200,000) had 33 storms; and the extreme damage category
(more than $200,000) consisted of 58 storms. The parameters associated with the minor, moderate, and extreme
categories were analyzed in detail.
The number of storms in each damage category that
qualified solely because of glaze damages, heavy snow,
or because of heavy snow and glaze combined are shown
in table 10. One minor damaging storm occurred with no
glaze and less than 6 inches of snow. There is a relationship between the degree of damage and relative freTable 10. Number of Damaging Storms for
Different Definitions
Number of damaging storms from given conditions
Heavy
Glaze and
heavy snow
Total
snow only
Damage
category
Glaze only
Minor
Modernte
Extreme
18
5
11
46
17
24
17
11
23
81
33
58
15
quency of the storms having both heavy snow and damaging glaze. These combined storms represented 21 percent
of 81 minor damage storms, 33 percent of the moderate
damage class, and 40 percent of extremely damaging
storms.
The monthly frequencies of damaging storms (table
11) show that many of those in the minor class occurred
in February and March, whereas the largest number in
the moderate class was in December, and that in the
extreme class was in January. Expressing the total storms
with damage as a percentage of all storms shows that 57
percent of the 304 storms were damaging, and 19 percent
of all storms produced extreme damage. Sixty percent
or more of the storms in November, February, and March
were damaging. However, a relatively large percentage
of the January and April storms were extremely damaging.
The number of storms per decade in each damage
category is shown in table 12. The low total number
(20) for the 1900-1910 period resulted from many unknown and probably minor damaging storms in that
period. Relatively few ‘badly damaging’ (moderate and
extreme) storms occurred in the 1931-1940 decade and
many occurred in the 1900-1910 and 1911-1920 periods.
The greatest number of extremely damaging storms in
one year was three, occurring in 1900, 1909, and 1912.
The greatest number of badly damaging storms in one
year was five in 1909, 1950, and 1951, and the greatest
number in one winter season was six in the 1951-1952
winter. In the 58-day period beginning on 6 November
1951 and ending 2 January 1952, there were three extremely damaging storms (on 6-7 November, 25 December, and l-2 January) and three moderately damaging
storms (on 14 December, 17-18 December, and 20-21
December). The greatest number in one month was four
in December 1951 and four in January 1918 (two extreme
and two moderate).
Table 13. Various Characteristics of Damaging Storms
Damage
category
Minor
Moderate
Extreme
Median
point
duration
(hours)
13.1
16.0
18.6
Median
percent
of Illinois
with glaze
in storm
Minor
Moderate
Extreme
2
4
9
Preferred
3-hour
period
of storm
initiations
Median
length
(miles)
6-inch snowfall cores
0900-1200
1000-1300
0800-1100
65
72
83
Median
width
(miles)
25
30
40
30°-sector
with most
orientations
236-265
231- 260
226-255
Average areal extent of snowfall
(square miles)
1 inch
1-6 inches
32,300
35,600
36,400
23,600
22,400
22,900
6-10 inches
7,500
9,800
9,300
>10 inches
1,200
3,400
4,200
The median point durations of the storms that produced moderate or extreme damages (table 13) were
considerably greater than those for all storms (table 5)
or those that produced minor damage. The time of
damaging-storm initiation in Illinois showed very little
difference for the different degrees of damage.
As shown in table 13, the median lengths and median
widths of the 6-inch snowfall cores increased with degree
of damage. The orientations of the cores in extremely
damaging storms were generally SW-NE, whereas those
with moderate and minor damages were generally WSWENE.
The areal extent of snowfall and glaze (table 13) also
increased with degree of damage. Two percent of Illinois had glaze with minor damage storms, as compared
with 9 percent with the extreme damage storms. The
average area with 1 inch or more of snowfall was 32,300
square miles for the minor damage storms, increasing to
36,400 square miles for storms producing extreme damage. In all three damage classes, the two state sections
where storm cores were most frequent were the northwest and northeast, and the two sections of lowest frequency were the southwest and southeast.
Table 11. Frequency of Damaging Storms per Month
Damage
category
Number of storms
Feb
Mar
Jan
Total
24
5
12
1
2
4
82
33
58
38
41
7
173
53
60
66
50
57
25
17
19
29
19
Dec
Minor
Moderate
Extreme
6
3
2
16
9
9
15
7
20
20
7
11
Total
11
34
42
61
52
11
14
Total with damage
as percent of
all storms
Extreme-damage
storms as percent
of all storms
Extremely Damaging Storms
Apr
Nov
Table 12. Number of Damaging Storms per Decade
Decade
1900-1910
1911-1920
1921-1930
1931-1940
1941-1950
1951-1960
16
Minor
2
12
12
19
17
20
Moderate
4
8
7
3
6
5
Extreme
14
12
9
6
7
10
Total
20
32
28
28
30
35
The 27 most damaging storms in the 1910-1960 period
are listed in table 14. Storms in the 1900-1909 period
were not included because less specific data on their
damages could be obtained. However, it is likely that the
27-28 February 1900 storm was one of the worst in the
61-year period. At the time of this 1900 storm, the
Weather Bureau records listed it as the worst storm in
Illinois since those of 1831, which was the worst winter
in Illinois since the white man settled in the area that
was to become Illinois.
The known dollar losses from physical damage for
each storm shown in table 14 were adjusted to the 1960
dollar value, using price indices set to a 1910-1914 base.8
This allowed comparison of losses and the assignment of
ranks. The indirect losses resulting from these storms
Table 14. Damages in 27 Worst Winter Storms,
1910-1960
Storm date
20-21 Feb 1912
21 Feb 1913
6-7 Jan 1918
11-12 Jan 1918
16-17 Apr 1921
11-12 Mar 1923
17-19 Dec 1924
30-31 Mar 1926
18-19 Dec 1929
25-26 Mar 1930
1-2 Mar 1932
7-8 Jan 1937
6-8 Apr 1938
29-30 Jan 1939
29-30 Jan 1947
31 Dec 1947–
1 Jan 1948
10 Jan 1949
17-18 Jan 1949
13 Feb 1950
9-10 Apr 1950
6-7 Nov 1951
25 Dec 1951
21-22 Mar 1955
8 Dec 1956
20-22 Jan 1959
9-10 Feb 1960
19-20 Dec 1960
Dollars of loss,
adjusted to
1960 value
Number
dead
Number
injured
Storm
rank
U
U
U
U
U
U
M
M
U
U
U
M
U
46
1
10
22
17
2
15
9
1
6
8
7
18
11
16
14
21
2,000,000
600,000
1,240,000
4,960,000
1,413,500
2,016,000
12,400,000
3,690,000
2,520,000
2,580,000
1,233,840
1,971,000
1,370,000
1,518,000
894,240
4
U
U
U
U
0
2
U
U
U
U
0
0
3
2
3,930,000
0
U
4
990,000
495,000
4,270,000
1,830,000
575,000
575,000
3,890,000
540,000
1,545,000
350,000
1,000,000
1
0
2
4
U
U
7
0
U
7
U
M
M
M
U
U
U
U
M
U
21
U
20
26
3
12
23
24
5
25
13
27
19
° U = unknown number, M = many but exact number not available
are not known, but the 8:1 ratio (table 9) between indirect and direct losses in the January 1967 storm suggests
that indirect losses would easily exceed the known direct
damages shown in table 14.
The patterns of the four most damaging storms are
portrayed in figure 14. The most damaging storm created
more than $12 million in property damages, killed 2, and
injured many. This was a glaze-only storm, with the
highest point snowfalls under 5 inches (figure 14a). The
direct loss in this storm was about 50 percent of that in
the 26-27 January 1967 storm, further revealing that the
1967 storm was the most damaging winter storm in Illinois in the period since 1900.
The second-ranked storm in the 1910-1960 period was
the opposite of the 1924 storm, being basically a snowonly storm with no glaze in Illinois. Although point snowfalls did not exceed 13 inches (figure 14b), much of the
state experienced high winds that produced blizzard
conditions and excessive drifting, and the entire state had
more than 1 inch of snow. The damaging conditions of
this storm persisted for 20 days after the storm.
The third-ranked storm (figure 14c) was a combination
snow and glaze storm, and was confined to the northern
third of Illinois. The damages in the fourth-ranked storm
(figure 14d) primarily came from exceptional glaze
loading and damages in northeastern Illinois (see table
28).
METEOROLOGICAL CONDITIONS WITH SEVERE WINTER STORMS
Historical surface weather maps prepared by the
U. S. Weather Bureau were the main source of data for
selecting the meteorological conditions which produced
the severe winter storms in Illinois between the months
of November and April for the 61-year study period.
Data published in issues of the Monthly Weather Review
were also utilized to supplement missing weather maps.
Storm
Classification
Because of its physical location, Illinois comes under
the influence of at least two principal storm tracks during
each cold season month (November through April).
During January and February, the state is subjected to
three principal storm tracks. 10 Examination of weather
conditions with the 304 storms indicated that with few
exceptions, severe winter storms in Illinois were associated with transitory cyclones. Cyclones which led to the
lllinois storms were classified according to the place of
origin, to the place of major intensification or re-formation, and to their route of movement.
This rather broad classification was necessary because
the available weather data prior to 1935 made it extremely difficult to determine whether the observed
cyclone was an original development or if it was an old
system that had temporarily lost its surface identity and
then moved into an area favorable for re-formation.
Almost all cyclogeneses in middle and high latitudes occur in connection with fronts. 11 The initial movement of
a cyclone is normally ill-defined. Only after the disturbance has intensified and moved out of its immediate
cyclogenic region does it attain a regular path. It tends
to follow over locally warm surfaces, such as inland
bodies of water. 10
Areas of favorable cyclogenesis initially were divided
into two localities, the lee of the North American Rockies
and the Gulf of Mexico. Because of the great latitudinal
extent of the North American Rockies, this area was
separated into two regions of frequent development,
Alberta and Colorado. All of these developmental regions
have the necessary physical features to maintain a quasistationary surface front. In Alberta and Colorado, adiabatic warming of air descending the mountains sustains
the lee-of-the-mountain fronts, 12, 13 and the warm water
surfaces in the Gulf of Mexico during the winter maintain a frontal boundary along the Gulf Coast. 14
In addition, all three areas are subjected to the penetration of arctic and polar air masses. Each of the three
areas lies in the path of cold core systems which move
17
a. Rank 1, 17-19 December 1924
c. Rank 3, 13 February 1950
b. Rank 2, 11-12 January 1918
d. Rank 4, 31 December 19471 January 1948
Figure 14. Patterns of snow, glaze, and other conditions associated with four most damaging storms in 1910-1960
18
Figure 15. Schematic of weather (cyclonic) types related
to severe storms in lllinois
out of the Pacific and may or may not be recognizable
from surface data by the time they reach their respective
cyclogenic areas. All three areas lie in the general vicinity
of normal fields of convergence during the winter
months. 15
Weather Types
Alberta cyclones, labeled as the type 1 storm producer,
were defined as those lee-of-the-mountain cyclonic developments which formed north of the U. S.-Canadian
border and moved easterly into the Great Lakes Basin
(figure 15). The border is a good dividing line, since it
is an area of minimal cyclogenesis. These Alberta cyclones produced 34 of the Illinois winter storms, or 11
percent of the total.
The 182 cyclonic developments in the Colorado area
that led to severe winter storms were separated into
three storm-producing types according to the route
traveled in their easterly migration. Type 2, the most
frequent severe winter storm producer in Illinois, formed
either in southeastern Colorado or in the adjacent Great
Plains area. This type tends to move into the junction of
the Mississippi-Ohio River Basins, and thence northeastward into the Great Lakes Basin and beyond (figure 15).
The route is parallel to and to the south of the preferred
winter position of the polar front which is oriented
WSW-ENE across central Illinois. 29 These type 2 cyclones
moved either continuously at a rapid rate or intermittently with stagnation. While stagnant, they received
additional intensification in the Ohio Valley or the Great
Lakes Basin. There were 137 storms, or 45 percent of the
total number, with type 2 development.
Type 3 was defined as that cyclone which appeared in
the northern Colorado and adjacent Great Plains area
and moved northward, sometimes as far as the Canadian
border. It then migrated into the Great Lakes Basin,
passing north of the Ohio River system. This type pro-
duced 23 storms, or 8 percent of the total. Type 4 was
defined by the cyclones that originated in either northern or southern Colorado and migrated eastward along
a route south of the Ohio River system. This type related
to 22 storms, or 7 percent of the total.
The Texas-West Gulf cyclone, labeled as type 5, departed from the Gulf of Mexico and took any one of
several paths northward. The type 5 cyclones produced
74 storms, or 24 percent of all the severe winter storms.
A few of the Alberta cyclones moved down the frontal
surface in the lee of the Rocky Mountains, crossing the
Colorado cyclogenic region, and on into the West Gulf
area. Here they received intensification and moved northward into a position where they produced a severe winter storm in Illinois. Such a system was classified as type
5, a Texas-West Gulf cyclone. Many other systems moved
directly eastward from the Pacific Ocean, but these were
classified into one of the five source-route types according to the route and area which last contributed to its
re-formation or intensification.
Although the vast majority of severe winter storms in
Illinois were produced by these five types of transitory
cyclones, 14 were associated with certain other synoptic
weather conditions. In two cases, frontal passages alone
produced the frozen precipitation. Five storms resulted
when the cyclone appeared to form in the Great Lakes
Basin. Four other severe storms resulted from migratory
cyclones that moved in a northerly direction west of the
Mississippi River and into the westernmost part of the
Great Lakes Basin. In three other instances, wave centers
associated with similar transitory cyclones moved diagonally across Illinois into the Lakes region. These infrequent conditions were grouped and considered as weather
type 6 in further analyses of the storms.
Climatological analysis of cyclone frequency in the
November-April period allows an estimation of the percentage of all cold-season cyclones that result in severe
winter storms in Illinois. In table 15, types 2, 3, and 4,
which originate in Colorado, have been grouped for
comparison with readily available statistics on the frequency of all Colorado cyclones. All cyclonic frequencies
were set to a base period of 20 years for ease of comparison. The frequency of Alberta cyclones was determined
from a 20-year sample 16 ; the 20-year frequency of ColoTable 15. Comparison of the Frequency of Cold-Season
Cyclones with the Number Producing Severe Winter Storms
in Illinois
Cyclone
source area
Alberta
(type 1)
Colorado
(types 2, 3, 4)
Texas-West Gulf
(type 5)
Total number
of cyclones
developed
in average
20-yr period
Number
of severe
storms produced
in Illinois
in average
20-yr period
Percent of
total cyclones
leading to
Illinois storms
487
11
2+
640
61
10—
304
25
8+
19
rado cyclones was extrapolated from data in a 3-year
sampling period 13; the value for the Texas-West Gulf
cyclones was calculated from a 40-year sample. 14
The percentages shown in table 15 reveal that only a
relatively small percentage of all cyclones formed, regenerated, or intensified in the three source areas actually
lead to severe winter storms in Illinois. The percentages
do show that cyclones from the Alberta source region
are considerably less frequent producers of severe winter
storms than are those in the other two regions.
Detailed Synoptic Analysis of January 1967 Storm
The greatest frequency of severe winter storms in
Illinois was produced by the type 2 cyclone. An example
of this type of storm occurred in Illinois on 26-27 January
1967, the storm previously described as very damaging
(see table 9). A detailed chronological synoptic discussion of this storm 17 is presented to illustrate more clearly
the upper- and lower-level atmospheric conditions with
this prevalent storm type. This particular storm very
closely followed the stages of winter storm development
described by Petterssen 11 : “The development at sea level
was preceded by a forward march and intensification of
an upper cold trough and commenced at the time when
the area of positive vorticity advection in advance of the
upper trough became superimposed upon the sea level
frontal zone.”
The description of events which produced this severe
winter storm begins with conditions at 0000 CST on 25
January 1967, when a continental arctic air mass entered
the United States in eastern Montana. A deep wave
moved over the Great Lakes Basin as a continental polar
air mass, which had formed in the Intermountain region
and moved across the Great Plains area. The stationary
front continued to exist along the lee of the Rockies while
maritime tropical air over the central and eastern United
States produced unseasonably warm temperatures. A
very small wave formed on the stationary front in southeastern Colorado in the general vicinity of the point
juncture of the three air masses cited above (figure 16).
A 500-mb trough located over Nevada, with its associated cyclonic vorticity center having a value of 14, moved
east at 34 knots.
By 0000 CST on 26 January (the start of the storm in
Illinois), the maritime polar air mass extended over the
entire western United States, and was separated from the
strong continental arctic air mass over the Midwest and
Great Plains (as far south as Oklahoma) by a frontal
zone extending along the lee of the Rockies and across
the central United States. Maritime tropical air dominated the remaining eastern and southeastern portions
of the United States, The point juncture of the three air
masses was the wave center with a central pressure of
1007.0 mb, and it was located in east-central Oklahoma
(figure 16). The warm front portion of the wave extended
up the Ohio Valley and into Canada.
20
Figure 16. Positions of significant synoptic weather features
during 25-28 January 1967
The 500-mb chart at 0000 CST on 26 January showed
a deep trough over the Great Plains located about 240
miles west of the surface wave center. An extensive area
of snow reached northward over the Great Plains to
North Dakota and eastward into Missouri and Iowa. Rain
and rainshowers continued to occur along the warm
frontal zone in the Midwest. The 1000-mb isobaric contours over Lake Michigan were oriented northeast so that
conditions in northeastern Illinois were favorable for
lake-effect (added warming and moisture) storms18 The
500-mb center of cyclonic vorticity with a central value
that had increased to 16 was being advected ahead of
the trough. Freezing rain in Illinois began shortly after
0000 CST on 26 January, with snow beginning between
0600 and 0900 in the northern areas. Precipitation in
excess of 2 inches fell during the next 24 hours in parts
of central and northern Illinois.
The wave center moved up the Ohio and was centered
near Louisville, Kentucky, at 0000 CST on 27 January
( near the end of the storm in Illinois ). The wave cyclone
dominated the entire central and eastern United States.
The cold front portion of the wave extended south
through Mobile and was preceded by a squall line. The
500-mb chart showed a closed low in the trough now
centered on the eastern Arkansas-Missouri border. The
surface wave center pressure had decreased to 997.8 mb,
and it continued to be the focal point of the three air
masses (figure 16). The center of positive vorticity which
had increased to 18 continued to be advected ahead of
the upper trough. The storm in various parts of Illinois
ended shortly after 0000 CST on 27 January. By 0000
CST on 28 January, the wave had occluded and deepened
in the vicinity of Lake Erie.
Frequency of Weather Types
The number of weather types in each month is shown
in table 16. The type 2 storm-producing condition prevailed in all months, although type 5 (Texas-West Gulf)
was relatively frequent in February. Type 1 (Alberta)
was relatively frequent in January.
Table 16. Monthly Number of Weather Types with
Severe Winter Storms
Type
OctNov
Dec
Jan
Feb
Mar
3
9
0
0
5
2
7
28
5
7
16
2
13
36
1
7
17
5
5
24
6
5
21
3
5
31
8
3
13
2
1
2
3
4
5
6
AprMay
1
9
3
0
2
0
Total
34
137
23
22
74
14
Table 17. Number of Weather Types Associated
with Different Storm Types
Weather
type
1
2
3
4
5
6
Glaze-only,
no heavy snow
Number
Perccnt
of storms
of total
3
17
3
1
6
4
9
12
13
5
8
29
Henvy snow,
plus glaze
Percent
Number
of storms
of total
Heavy snow,
no glaze
Number
Percent
of total
of storms
27
89
13
19
55
9
80
65
57
86
74
64
4
31
7
2
13
1
11
23
30
9
18
7
The frequency of different weather types with the
three basically different types of severe winter storms is
shown in table 17, and the percentages allow evaluation
of the importance of each type. Weather types 1 and 4
were more frequently associated with the heavy snow
(no glaze) type of winter storms than were the other
weather types. Weather types 2 and 3 were alike in that
they both had a distinct tendency for a more frequent
association with the glaze-only and the snow-plus-glaze
storms than did the other weather types. Weather type
6 had a much higher percentage of association with the
glaze-only storms than did types 2 and 3, but type 6 had
little association with the snow-plus-glaze storm type.
The frequency of the six weather types with all 304
storms and with the three main categories of damaging
storms is shown in table 18. The frequencies for each
weather type were expressed as percentages of the total
per type to allow evaluation of the relative importance
of each weather type for each damage class. The percentages for types 4 and 2 show the relative importance for
their associated storms to produce extreme damage.
Weather types 1 and 6 are more important than other
types in producing storms of moderate damage. The 5
and 3 weather types show a stronger tendency for assoTable 18. Frequency of Weather Types Associated
with Storms with Different Degrees of Damage
Weather
type
1
2
3
4
5
6
Minor
Extreme
Moderate
damage
damage
damage
All storms*
Percent
Percent
Percent
Percent
Number of total Numler of total Number of total Numler of total
4
31
5
4
11
3
24
38
28
40
31
30
5
15
4
2
4
3
29
18
22
20
11
30
8
36
9
4
21
4
47
44
50
40
58
40
*Includes storms with no damages or unknown damages
34
137
23
22
74
14
11
45
8
7
24
5
ciation with storms producing only minor damages than
do the other types. The type 2 and 4 weather conditions
which most often led to storms producing extreme damages in Illinois are both conditions with Colorado lows
that have either an east track or south track across the
Midwest.
Effect of Lake Michigan on Severe Winter Storms
Lake Michigan occasionally plays a role in the occurrence or intensification of severe winter storms in northeastern Illinois. In these instances, the migratory cyclone
which produces the severe storm has a trajectory such as
to orient its isobaric contours cyclonically from southeast
through north. Such an orientation occurs when the
wave center passes south and east of Illinois (types 2
and 5). This route produces low-level winds from east
through northwest, but the more northerly winds have
the longest fetch across Lake Michigan. The warmer
wintertime open water adds 1) a significant amount of
heat to the cold air mass passing over the lake and 2)
moisture to relatively dry air, both of which can act to
enhance snowfall in northern Illinois.
To examine for the possible effect of Lake Michigan on
the occurrence or intensification of severe winter storms,
the mean isobaric orientations at the 1000-mb level over
the lake during the storm period were determined for
all 269 storms having more than 6 inches of snow. Nine
orientations and rapidly changing cyclonic and anticyclonic configurations were found. Overlake flow towards
Illinois (orientations of NNE, NE, ENE, and E) occurred
in 164 storms, cyclonic in 27, anticyclonic in 8. Overlake
orientations which should not affect Illinois (NW, NNW,
N, ESE, and SE) occurred with 105 storms. In general,
the gross effect of moisture infIux from the Atlantic
Ocean and the Gulf of Mexico predominated and obscured any significant lake effect on most of the storms,
Certain indications of lake effect are shown in the
area1 distribution of 6-inch snowfall cores (figure 17).
The frequency pattern for the storms with the favorable
overlake isobaric orientations (figure 17a) showed much
more latitudinal variation with the two northern sections
experiencing cores 2 to 3 times more often than did the
southernmost districts of Illinois. With no overlake flow,
the northern districts had only 1.2 to 1.5 times more
heavy snowfall cores.
The storms with overlake flow represented 54 percent
of all storms. There were 65 storms with a maximum
point snowfall of 12 inches or more, and 52 of these (70
percent) occurred with overlake isobaric orientations.
These facts suggest an intensification of snowfall in Illinois with overlake flow.
A total of 31 winter storms had snowfall maximums
in close proximity to the lake and restricted to the nearlake area. These patterns appeared to be a direct result of
lake effects. 18 Ten of these storms had only one maximum,
21
a. Overlake flow
(orientations NNE, NE, ENE, E)
b. No overlake flow
(orientations NW, NNW, N, ESE, SE)
Figure 17. Number of times 6-inch snow cores occurred in each district (1900-1960) with different
1000-mb isobaric orientations over Lake Michgan
that at or near the lake, and would not have qualified
as severe winter storms without the lake effects. The 21
others had a lake-produced maximum and one or more
cores elsewhere in Illinois. Thirteen of these storms reflecting strong lake effects occurred in November and
December, when the lake water is relatively warm, 19
providing a source of heat as well as moisture. The lakeeffect snowstorms assume importance in Illinois because
they occur in the complex Chicago urban area where
6 inches or more snowfall in 48 hours or less is a relatively more serious problem than it is elsewhere in
Illinois.
PATTERNS OF SELECTED WINTER STORMS
Certain unusual winter storms or storms occurring in
unusual circumstances have been portrayed in this section to illustrate their patterns. The patterns for the
earliest (28-30 October 1925) and latest (l-2 May 1929)
winter storms are shown in figures 18a and b. Neither
storm produced any glaze in Illinois nor even minor damages ( less than $1000). The May storm was not extensive
and had a very narrow path across Illinois. Figure 18c
illustrates the pattern of a storm oriented NW-SE across
Illinois, one of the three storms with such an orientation.
Figure 18d illustrates another infrequent type of winter
storm. In this ‘southern’ storm, the core is just barely
located in extreme southern Illinois.
Figure 19 presents six storm patterns illustrating lakeeffect heavy snowfalls in northeastern Illinois. Figures
22
19a, b, and c show 3 of the 10 lake-effect storms with a
single maximum in the state. In figure 19a, the maximum
is inland and southwest of the lake; in figure 19b, the
maximum is small and located on the lakeshore; and in
the third example, the maximum is inland and west of the
lake (figure 19c).
Figures 19d, e, and f are examples of the 21 storms
with a lake-effect maximum in addition to maximums
elsewhere in Illinois. These storms would have achieved
a severe classification without the lake-effect high, but
the lake did produce heavy snow in northeastern Illinois.
The shape of the high near the lake in figure 19d is
similar to that in figure 19b. The lake-effect maximum
shown in figure 19e exhibits still another pattern, with
a narrow ridge parallel to and extending along the entire
a. Earliest storm of season,
28-30 October 1925
c. Storm with NW-SE orientation,
27 November 1922
b. Latest storm of season,
1-2 May 1929
d. Storm centered in south,
2-3 February 1939
Figure 18. Examples of snowfall patterns for unusual storms
23
a. 12-13 December 1903
d. 29-30 March 1928
b. 10 December 1934
c. 26 November 1950
e. 17-19 January 1943
f. l-2 March 1947
Figure 19. Examples of storm snowfall patterns reflecting lake effects
lakeshore in Illinois. Figure 19f shows an isolated high
westward and inland from the lake, but it is not centered
in the same location as that of the high in figure 19c.
24
The location of the heavy snowfall center resulting from
flow over Lake Michigan varies and is dependent on the
orientation of the flow (length of overlake travel), wind
Snowfall, glaze, sleet, and blowing snow on
a. 11-12 March 1923
b. 20-21 March 1932
c. 9-10 February 1960
Thunderstorms, hail, damaging lightning and winds, precipitation,
excessive rainfall, and tornadoes on
d.
11-12 March 1923
e. 20-21 March 1932
f. 9-10 February 1960
Figure 20. Patterns of three freak winter storms
25
a. 27-28 February 1900
b. 6-8 February 1933
c. 20-22 January 1959
Figure 21. Three winter storms producing the heaviest point snowfalls in 1900-1960 period
speed, and relative differences between the temperature
and humidity of the air and that of the lake and land just
downwind of the lake. 18
Three other severe winter storms were quite unusual
because they each combined the occurrence of heavy
snow, high winds, damaging-widespread glaze, sleet,
thundérstorms, hailstorms, and tornadoes in Illinois. 20
Their patterns (figure 20) reveal a general latitudinal
distribution, with the severe weather events common to
the warm season largely in central and southern Illinois.
However, many persons in the 1932 storm had the unusual experience of witnessing a lightning storm while
heavy snow was falling. All three storms caused sufficient damage to be in the extreme damage class, and
two (1923 and 1960 storms) were ranked as the 9th and
19th worst storms respectively (table 14). All three
storms were produced by weather type 2, the most common type. Studies of midwestern snowstorms have shown
that thunderstorms embedded in a snowfall zone could
increase local falls by 2 inches or more if the thunderstorm persisted for 3 hours. 21
The patterns of the three storms producing the highest
point snowfalls in the 1900-1960 period are shown in
figure 21. The 37.8-inch maximum in western Illinois on
27-28 February 1900 was considerably greater than that
in the storm with the second highest value, 22.1 inches
in the 6-8 February 1933 storm. (As a comparison, the
more recent 26-27 January 1967 storm had a point maxi26
mum of 26.0 inches, at Peotone just south of Chicago,
which would place it second after the 1900 storm.) The
point maximum in the 1959 storm in figure 21 was 21.5
inches. The 1900 storm had two areas with snowfall in
excess of 25 inches and was extremely damaging. The
1933 storm caused only minor damages, but the 1959
storm caused extensive damages, being ranked as 13th.
The greatest number of badly (moderate or extreme)
damaging storms in one month was four in January 1918.
The patterns of these storms appear in figure 22. The
storms on 6-7 and 11-12 January were both classed as
extremely damaging, ranking as the 17th and 2nd most
damaging in the 1910-1960 period, respectively (table
14). The storms on 14-15 and 26-28 January produced
damages in the moderate class. All four storms were
basically ‘snow-only’ storms, and only the 6-7 January
storm rated also as a glaze storm. At the end of January
1918, the snow on the ground in Illinois was 35 inches
in northern Illinois, 10 to 20 inches in the central districts,
and 7 to 10 inches in southern Illinois.
The greatest number of badly damaging storms in Illinois in a 2-month period occurred in 1951-1952. Six
storms occurred in a 58-day period beginning with a
storm on 6-7 November 1951. The patterns of these six
storms are shown in figure 23. The storms on 6-7 November, 25 December, and l-2 January all were extremely
damaging and the other three were moderately damaging.
A minor damaging storm occurred on 17-18 January also.
a. 6-7 January 1918
c. 14-15 January 1918
b. 11-12 January 1918
d. 26-28 January 1918
Figure 22. Severe winter storms in January 1918
27
b. 14 December 1951
a. 6-7 November 1951
d. 20-21 December 1951
e. 25 December 1951
Figure 23. Series of six damaging storms in 58-day period
28
c. 17-18 December 1951
f. l-2 January 1952
Part 2. Climatography of Glaze Storms
Glaze-producing (ice) storms in the 1900-1960 period
were selected for separate study because their damageproducing capability and related design problems are
significantly different from those associated with snow.
In the 61-year period there were 92 glaze storms defined
either by the occurrence of glaze damage or by occurrence of glaze over at least 10 percent of Illinois (as determined by station density), and these 92 represent 30
percent of the total storms. Thirty-four of these glaze
TEMPORAL
storms were solely glaze storms and occurred without any
6-inch snowfalls in Illinois, whereas the other 58 were
associated with heavy snowfall. Sixty-nine of the remaining 212 storms had some glaze reports but no glaze damages and too few point reports of glaze to qualify as
widespread glaze (10 percent of Illinois; see table 7).
The study of glaze storms was restricted to the 92 storms
qualifying as damaging, widespread, or both.
CHARACTERISTICS
Annual Frequencies
The number of severe winter glaze storms in Illinois
during an average 10-year period is 15, but there have
been as many as 28 storms in a 10-year period, (19441953), and as few as 9 (1902-1911). The greatest number of glaze-producing storms in one year was 6 (1951);
in two years, 9 (1950-1951); in three years, 10 (19501952); and in five years, 15 (1948-1952). There were
no glaze storms in 12 of the years during 1900-1960.
Monthly Frequencies
The total number of storms in each month is shown in
table 19. The greatest numbers occurred in December
and January, and these represented slightly more than
one-third of all severe winter storms in these two months,
The number of winter storms in February, March, and
April that were glaze storms are quite comparable, 28
or 29 percent. Percentages in table 19 reveal that relatively few of the November winter storms produced
damaging glaze or widespread glaze.
could be determined for 85 of the 92 glaze storms. The
number of initiations in each of the six 4-hour periods of
the day is shown in table 20. The glaze-only storms
Table 20. Number of Glaze Storm Initiations
in 4-Hour Periods
Conditions
Glaze-only
(no heavy
snow)
Glaze with
heavy snow
Total
showed
period,
almost
periods
0000–
0400
0401–
0800
5
5
Time, CST
0801–
1201–
1200
1600
10
6
1601–
2000
3
2001–
2400
3
8
7
13
11
11
3
13
12
23
17
14
6
a preference for initiation in the 0801-1200 CST
whereas the glaze-plus-heavy-snow storms showed
equal numbers of initiations in the three 4-hour
beginning after 0800.
Storm Point Durations
Table 19. Monthly Number of Glaze Storms
Conditions
Nov
Dec
Jan
Feb
Mar
Apr
Total
1
10
10
4
8
1
34
Glaze-only
(no heavy snow)
Glaze with
heavy snow
Total
Number of glaze
storms as a percent
of all winter storms
2
12
17
14
10
3
58
3
22
27
18
18
4
92
17
35
34
28
29
29
30
Time of Occurrence
The time that each glaze storm initiated in Illinois
The average point storm duration could be determined
for 82 of the 92 glaze storms. The values ranged from a
low of 2 hours to a high of 48 hours. The median point
duration of the glaze-only storms was 12.0 hours, and
that of the glaze with heavy snow was 13.1 hours. Both
values are less than the 14.2-hour median determined for
all storms (table 5). Only 14 of the 82 glaze storms with
discernible durations had point durations greater than
20 hours, and 11 of these were storms with glaze and
heavy snow. There was little difference between the
median durations determined for each month.
29
AREAL-GEOGRAPHICAL CHARACTERISTICS
Geographical Distribution of Glaze Areas
The geographical
recorded as being in
central, central and
central, and south).
over all three areas
number, 18, occurred
location (figure 24) of glaze was
the north, central, south, north and
south, and all three areas (north,
Twenty-one glaze storms occurred
(table 21), and the second largest
in only the central area.
Table 21. Number of Glaze Storms in Various
Areas of Illinois
Areas
North
Central
South
North and
central
Central and
south
North, central,
and south
Nov
Dec
Jan
Feb
Mar
Apr
Total
1
1
0
1
5
4
1
5
2
2
2
4
6
3
2
2
2
0
13
18
12
0
3
5
4
4
0
16
0
2
7
2
1
0
12
1
7
7
4
2
0
21
In December and January, the 3-area occurrences predominate, with a considerable number of occurrences for
the central area and for the southern area. In March,
April, and November, glaze occurrences were predominantly in northern and central Illinois.
The monthly and total number of times a glaze region
occurred in each area are depicted in figure 24. The
central area totals equal or exceed those of the other
areas in all months except March. Southern totals rank
second in December, January, and February; northern
experiences rank second in February (ties south), and
first in March, April (ties central), and November (ties
central). These monthly values show a northward shift
with time in severe glaze occurrences during the winter
season. The winter totals for each of the three areas
(figure 24) indicate that the greatest frequency was in
central Illinois with frequencies in southern and northern
Illinois being nearly equal.
Other maps of the areal distribution of glaze storms,
as defined by different sets of data, 7 are shown in figure
25. Three of these maps show a latitudinal distribution
with the most occurrences somewhere in northern Illinois
and the fewest in southern Illinois. However, the pattern
on figure 25d, based on railroad measurements of ice on
wires, shows a central Illinois maximum which agrees
with findings from this study (figure 24). The differences between the patterns in figures 25a, b, and c and
that in figure 24 are considered to result from the different sampling periods. The 61-year period should provide a more representative measure than the 9- to 28-year
periods.
30
Figure 24. Number of times glaze occurred in each area on days
with glaze storms, 1900-1960
Areal Extent of Glaze Areas
The percent of Illinois experiencing glaze was determined for each of the 92 storms with widespread glaze,
or damaging glaze, or both. The percentages were determined from the number of stations reporting glaze by
allowing each of these to represent a given portion of
Illinois. This portion varied depending on the total number of weather stations existing in Illinois in the year of
the storm (see page 5). The storm percentage values
were used to calculate monthly average percentage values
(table 22). The averages reveal that glaze areas were
Table 22. Areal Extent of Glaze in Glaze Storms
Storm
area
Average
Maximum
Minimum
Nov
Percent of Illinois with glaze per storm
Feb
Apr
Dec
Jan
Mar
6.3
11
1
18.5
46
4
20.7
43
2
11.6
32
1
8.7
28
1
10.7
25
2
All
15.1
46
1
most widespread in January and December, and considerably less widespread in the other four months.
The six storms in the 1900-1960 period with the most
extensive glaze areas are shown in figure 26. These patterns reveal the shape and size of glaze areas when glaze
a.
c.
Number of glaze storms in 1929-1937,
Association of American Railroads data
Glaze storms causing damage to
telephone wires and poles, 1918-1925
d.
b.
Total severe glaze storms in 1925-1953
from U. S. Weather Bureau data
Storms with 0.25” or more radial thickness of ice on railroad wires, 1929-1937
Figure 25. Patterns of glaze storm frequencies as depicted by data from different agencies and for different base periods
31
a. 13-14 December 1937
d. 31 January 1908
b. 11 January 1905
e. 25-27 December 1904
c. 23-25 December 1945
f.
17-18 January 1949
Figure 26. Six storms in 1900-1960 with most extensive glaze areas, and percent of state covered by each
32
is extensive in Illinois. All produced damaging glaze, and
were ranked as either moderately or extremely damaging
storms, as based on total damages. The 1949 storm in
southern Illinois produced considerable tree damage on
the valuable forests and fruit orchards of that area. 22
Motion of Storms
The motion of the glaze-producing storms across Illinois could be determined for 86 storms. The number
sorted into 30-degree sectors (table 23) revealed that
the maximum number of storm motions were in the
241-270 degree sector (WSW), with minor maximums in
the 151-180 (SSE), 211-240 (SW), and 301-330 (NW)
degree sectors. The glaze-only storms exhibited a maximum of motion from the SW (211-240 degrees), whereas
the glaze with heavy snow moved most often from
the WSW.
Table 23. Directions from which Glaze Storms Moved
Sorted into 30-Degree Sectors
30-degree
sector
121-150
151-180
181-210
211-240
241-270
271-300
301-330
331-359
Total
Glaze-only
Number of storms
Glaze with
heavy snowfall
Total
1
6
3
12
5
3
4
0
34
4
4
2
3
26
3
5
5
52
5
10
5
15
31
6
9
5
86
Orientation of Glaze Areas
The general shape of the glaze areas could be determined for 62 of the glaze storms. These areas tended
to be elongated. The long axis of 50 of these areas (80
percent) had orientations varying from 225 to 260 degrees (SW-NE through WSW-ENE). Eight other glaze
area orientations were in the 261-280 degree sector, and
four were in the 180-224 degree sector.
METEOROLOGICAL CONDITIONS WITH GLAZE STORMS
Seven glaze storms (8 percent) occurred with Alberta
lows ( type 1), and 48 (52 percent) occurred with Colorado lows having a southerly track (type 2). Only 42
percent of the snow-only severe winter storms were produced by the type 2 conditions, revealing that weather
type 2 not only leads in glaze storm production but is
relatively more important in producing damaging or
widespread glaze than in producing 6-inch or greater
snowfalls. Ten of the 92 glaze storms were with type 3
cyclones (figure 15), 3 were type 4, 19 were type 5,
and 5 resulted from other weather conditions grouped
as type 6.
Riehl 23 indicated that most Illinois glaze storms occur
during the northeastward advance of a low center (types
2 and 5), and are quite uncommon with lows that move
eastward (type 4) or southeastward (types 1 and 3). He
defined two kinds of glaze storm situations: 1) a polar
front wave moving northeast along the eastern edge of a
continental polar outbreak, and 2) a broad southwesterly
current ascending over a sluggish body of cold air in the
eastern United States.
In weather types 2 and 5, a warm front frequently lies
across Illinois with the rain-producing maritime tropical
air overrunning the narrow wedge of cold polar air below
it, and this combination of circumstances is necessary for
glaze formation. In a synoptic analysis of 12 ice storms
at Springfield, Illinois, Bennett 7 found that seven storms
came with warm fronts, three with cold fronts, one with
no front, and one with an occlusion.
Another analysis of severe glaze storms in the Upper
Mississippi Valley showed that only 27 percent were from
the Texas-West Gulf area,7 which agrees with the 21 percent (19 out of 92) found for type 5 in Illinois. Bennett
also concluded that a major number of glaze storms in
the Illinois area were Colorado lows with a southward
track (type 2). His findings on storms produced by lows
from the west or northwest (types 1 and 3) showed that
20 percent were of these types. This agrees with the 19
percent for types 1 and 3 given in table 17, but disagrees
with Riehl’s conclusions. 23
33
Part 3. Design Information
Various point and area data on snowfall, glaze, and
related conditions were analyzed to provide 1) answers
to questions of structural design affected by these conditions and 2) information relating to various commercial
operations affected by these severe conditions. Selection
of the information presented was based largely on the
type of requests for information received, prior studies,
and on the obvious problem conditions.
The first phase of this design section concerns point
data. Included are point frequencies of heavy snowfall,
maximum point snowfalls on record, maximum annual
and storm snowfalls in Illinois, maximum point snow-
depth values, point glaze-day frequencies, maximum
point glaze accumulations, glaze with wind relationships,
frequency of days with snowfalls of different magnitudes,
and frequency of days with sleet.
The second phase of the design section concerns various areal-geographical conditions. Included are areadepth snowfall relationships for the severe winter storms,
models of snowstorms and glaze storms, regional probabilities of snow and glaze storms, temporal probabilities
for the occurrence of a storm after one has occurred, and
data on the temporal persistence of heavy snow and glaze
in the storm center after a storm.
POINT DATA
Point Frequencies of Heavy Snowfall
For the study of various point frequencies of shortterm, heavy snowfall, data from 23 stations (figure 27)
with 61-year records were analyzed. Point values for
clock-hour periods of 24 hours or less could not be
discerned at the 17 substations because their snowfall
data were recorded only once daily and at the same time
every day. Therefore, this analysis was based on fixed
2-day (calendar) totals, which in many instances do not
represent 48-hour maxima. However, at most stations at
least 30 percent of the 25 heaviest 2-day totals in the 61year period were values recorded for only one calendar
day with no snowfall on the days before or after it. The
2-day values were amounts that likely fell in less than
24 hours in many instances since median point severe
storm durations were less than 15 hours (table 5). A
similar analysis was based on maximum snowfall values
for 7-day (calendar) periods.
The highest 25 values for 2-day and 7-day periods at
each station were ranked and used to derive frequency
distribution curves. From these curves, values for various
return periods ranging from 2 to 50 years were ascertained. These values for each station were compared
with all other station values to determine which stations
had similar recurrence interval values. Five areas
(figure 27) of similar recurrence interval values were
found. The 2-day station values in each area were
averaged to determine the area mean values shown in
table 24; values determined for 7-day periods are shown
in table 25.
The 2-day area-average curves in figure 28 reveal that
area 3 has lower values than area 2 at the 2- and 5-year
intervals, but the area 3 values exceed those in area 2
at the less frequent recurrence intervals. Area 4 in the
south has higher expected values at intervals beyond 25
34
Figure 27. Data stations and areas of similar point snowfall
frequencies
years than has area 2, which is farther north. Areas 3,
4, and 5 all have 50-year values that are almost equal.
Area 1 in northern Illinois has much higher values at all
intervals than the other areas.
The distribution of maximum snowfall values for 7-day
periods shown in table 25 indicates that the 2- and 5-year
values of the five areas are considerably different, and
Table 24. Two-Day Maximum Snowfall Values for Selected
Recurrence Intervals in Five Areas of Illinois
Area
1(North)
2(West)
3(SW-central)
4(South)
5(Extreme
south)
Snowfall (inches) equalled or exceeded
at given interval
5-yr
10-yr
2-yr
20-yr
50-yr
30-yr
8.5
7.4
7.0
5.7
4.2
10.6
8.6
8.6
7.4
5.8
12.4
9.8
10.2
9.1
7.5
14.8
11.2
12.1
11.2
9.7
16.2
12.0
13.4
12.6
11.3
18.3
13.4
15.2
14.8
13.5
Table 25. Seven-Day Maximum Snowfall Values for Selected
Recurrence Intervals in Five Areas of Illinois
Area
1(North)
2(West)
3(SW-central)
4(South)
5(Extreme south)
Snowfall (inches) equalled or exceeded
at given interval
2-yr
5-yr
10-yr
20-yr
30-yr
50-yr
10.2
8.2
7.6
6.5
4.6
13.3
10.8
10.1
9.1
7.0
16.7
13.2
12.8
11.8
9.4
21.0
16.3
16.3
15.2
12.7
24.0
18.2
18.7
17.6
15.3
28.2
21.5
22.0
21.3
18.9
tions (figure 29) indicate that they occurred on 21 different dates, and only three 2-day periods produced a
maximum at more than 1 station. These three dates were
27-28 February 1900 (maximum at 3 stations), 18-19
February 1908 (maximum at 3 stations), and 12-13
January 1927 (maximum at 4 stations). The occurrence
of many different periods related to the 28 maximum
values indicates that extremely heavy snowfall zones in
Illinois are usually not widespread.
Nine of the 21 record snow periods occurred in February, six were in January, four in March, and two in
December. Thirteen of the 21 periods came in the 19001918 period, and only eight record point snowfall periods
occurred in the 42 years after 1918. The temporal distribution of severe winter storms (figure 3) did not indicate a high frequency of storms in the 1900-1918 period,
but there were many more extremely damaging storms
in the 1900-1920 period (table 12) than in any other 20year period after 1920.
Maximum Point Snowfalls per Storm and Year
The maximum point snowfalls recorded in the 269
severe winter storms with amounts of 6 inches or more
Figure 28. Area-mean recurrence interval curves for 2-day point
snowfall values
that all values for area 1 (north) are greater than those
in the other four areas. Areas 2, 3, and 4 have comparable
values for the 20-year and longer recurrence intervals.
Record High 2-Day Snowfall Values
The record high 2-day snowfall values in the 1900-1960
period at 28 stations with 61-year records are shown in
figure 29 along with their dates of occurrence. The highest value shown in the 61-year period was 26.0 inches at
Quincy on 27-28 February 1900. The highest 2-day value
at any Illinois station at any time in the 1900-1960 period
was 37.8 inches at Astoria on 27-28 February 1900 (figure
21), but it was not plotted in figure 29 since the Astoria
station was not one with a long-term operation.
Areas where relatively high maximums have occurred
include northern, western, northeastern, and southeastern
Illinois. Areas with relatively low (less than 14 inches)
maximums include southwestern and west-central Illinois.
The lowest maximum value was 11.8 inches at Moline.
The dates of maximum values shown for the 28 sta-
Figure 29. Maximum 2-day snowfall totals, in inches, and dates
of occurrence in 1900-1960
35
Table 27. Annual Maximum 2-Day Point Snowfall
Values in Illinois
Table 26. Maximum Point Snowfalls for Storms
Producing 6 Inches or More
Number of storms
OctNov
Dec
6.0–8.0
8.1–10.0
10.1–12.0
12.1–14.0
14.1–16.0
16.1–18.0
18.1–20.0
20.1–22.0
22.1–24.0
> 24.1
8
4
3
1
1
0
0
0
0
0
21
15
10
4
4
1
0
0
0
0
Totals
17
55
Snowfall
(inches)
Year
Feb
Mar
AprMay
Total
25
15
14
7
3
1
4
1
0
0
22
15
7
8
3
3
0
1
0
1
17
13
8
8
3
2
2
0
0
0
7
1
1
4
1
0
0
0
0
0
100
63
43
32
15
7
6
2
0
1
70
60
53
14
269
Jan
were sorted as shown in table 26. There were 31 storms
(upper 12 percent) with point values of 14.1 inches or
more. Most of these heavy values came with storms in
January, February, and March. Five of the nine storms
with the heaviest point amounts occurred in January,
further indicating the greater likelihood for heavy snow
in midwinter.
The highest 2-day point value anywhere in Illinois in
each year during the 1900-1960 period is shown in table
27. A value of 10 inches or more occurred in 52 of the
61 years. In 21 years the annual high was somewhere in
the northwest section, and in 14 years it was in the northeast. The annual high occurred only once in the central
and east-southeast districts.
Maximum Snow Depth
The snow depth (snow on the ground) data at the 28
stations (figure 29) with long records were analyzed to
determine the maximum depth recorded on any day in
the 1900-1960 period. Isodepth lines based on these
record values were drawn to derive a statewide pattern
(figure 30). This pattern shows that the lowest maximum
depths, less than 15 inches, have occurred in a corridor
across southern Illinois and in a small area in west-central
Illinois. Values in excess of 25 inches have occurred in
all of northern Illinois and in isolated areas in western
and eastern Illinois.
Assuming a 10:1 ratio of snow to water, 24 the 15-inch
maximum depths represent a weight of 7 pounds per
square foot and the 35-inch depths in extreme northern
Illinois are equal to a load of 18 pounds per square foot.
Maximum Glaze Accumulations
Because of the size, shape, and open exposure of most
communication and power service wires, they have a
greater collection efficiency and normally acquire thicknesses of glaze equal to or greater than that on most other
exterior objects. Bennett 7 reported that ice thickness on
highways is as great as that on wires about 80 percent of
the time. Therefore, many measurements of glaze in the
36
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
Date
27-28 Feb
2-3 Feb
15 Dec
12-13 Dec
25-26 Jan
7-8 Feb
26-27 Feb
15-16 Jan
18-19 Feb
12 Dec
16-17 Feb
5-6 Feb
16 Feb
26-27 Feb
12-13 Feb
24-25 Dec
2-3 Mar
7-8 Dec
6-7 Jan
6 Dec
5 Apr
16 Apr
18 Jan
11-12 Mar
20 Mar
28-29 Oct
30-31 Mar
12-13 Jan
17-18 Feb
18-19 Dec
25-26 Mar
6-7 Mar
20-21 Mar
6-7 Feb
10 Dec
22-23 Dec
16-17 Jan
13 Mar
7-8 Apr
30 Jan
13-14 Jan
10-11 Mar
5-6 Dec
17-18 Jan
10-11 Feb
24-25 Dec
8-9 Mar
26-27 Mar
1-2 Mar
17-18 Jan
7-8 Dec
20-21 Dec
2-3 Mar
1-2 Mar
2-3 Mar
21-22 Mar
11-12 Feb
9-10 Jan
20-22 Jan
20-22 Jan
9-10 Feb
Snowfall
(inches)
Location
State
section
37.8
14.0
8.0
13.5
15.0
11.2
14.0
10.5
18.0
15.0
18.0
14.3
16.5
11.7
15.0
16.0
9.0
17.0
19.0
7.0
12.0
14.0
6.3
11.0
12.0
6.5
19.0
19.0
13.0
12.0
19.0
15.0
14.0
22.0
12.8
9.0
11.0
11.5
15.7
18.3
13.3
12.2
11.1
12.0
12.5
11.0
9.5
15.0
13.8
8.0
15.2
19.0
13.0
10.0
11.8
9.0
11.0
12.0
11.5
21.6
15.1
Astoria
Wheaton
Astoria
Chicago Heights
Albion
Dixon
Mt. Vernon
Monmouth
Astoria
Freeport
Grafton
Chicago
Greenville
Galva
E. St. Louis
Quincy
Grafton
Cairo
Oregon
Freeport
La Harpe
Freeport
Kankakee
Freeport
Griggsville
McLeansboro
Dixon
Kankakee
Joliet
Danville
Chicago
Griggsville
Oregon & Sycamore
Joliet
Chicago
Morrison
La Harpe
Edwardsville
Chicago
Chicago
Monmouth
Freeport
Lincoln
Rockford
Rockford
Dixon
Mt. Carroll
Harrisburg
Rockford
Dixon
Wheaton
Galena
Mt. Carroll
Charleston
Chicago
Monmouth
Park Forest
Sycamore
Dixon
Rushville
Mt. Carroll
W
NE
W
NE
SE
NW
SE
W
W
NW
WSW
NE
WSW
NW
SW
W
WSW
SW
NW
NW
W
NW
E
NW
WSW
SE
NW
E
NE
E
NE
WSW
NW-NE
NE
NE
NW
W
WSW
NE
NE
W
NW
C
NW
NW
NW
NW
SE
NW
NW
NE
NW
NW
ESE
NE
W
NE
NE
NW
W
NW
various severe storms in Illinois have referred to the radial
thickness of ice on wires or to the weight of ice on wires.
In some instances, the relative difference between the
weight of tree branches and that of the ice they have
collected has been noted, since tree damage is a major
factor in glaze storm damages.
subsequent high winds in the hours and days following
the storm. 25 The winds furnish additional transverse
loading, causing more breakage of the heavily glazed
wires, poles, and trees. Much damage ensues when icecovered tree branches fall on houses and wires. 7
To obtain heavy glazing, it is necessary to have 1)
continuous, relatively heavy precipitation which is supercooled by falling from a warm (above freezing) raingeneration layer of air through 3000 to 4000 feet of air
with temperatures of 25 to 32 degrees F; and 2) moderately strong winds and cold surface objects that by evaporation and conduction can remove the latent heat of
fusion (resulting when drops impact) to produce a high
collection efficiency of the drops on surface objects.7
The rate of accumulation of ice on an unheated object
oriented normal to the wind is
I = V W 3600 (10 -4/2.54 ) ( E /100)
where I is the icing rate in inches per hour, V is the wind
speed in mph, W is the liquid water content in grams
per cubic meter, and E is the collection efficiency. 6
The Association of American Railroads made a 9-year
national study of glaze accumulations on their communication wires. 25 These data were summarized by
Bennett 7 according to the greatest point thickness measured in 60- by 60-mile squares, and these values are reproduced in figure 31. These data show that radial thickFigure 30. Maximum snow depth in inches, 1901-1960
Certain of these ice measurements recorded in some of
the most severe Illinois glaze storms are listed in table
28. These do not represent all the extremes, but are those
found in the records. However, they do serve as good
approximations of the possible extremes. The listings of
the large radial thicknesses show various locations
throughout Illinois, indicating that severe glazing has
occurred and can occur anywhere in Illinois.
In many of the severest glaze storms, the damage has
resulted from a combination of excessive ice loads and
Table 28. Measures of Glazing in Various Severe
Winter Storms
Storm date
2-4 Feb 1883
20 Mar 1912
21 Feb 1913
11-12 Mar 1923
17-19 Dec 1924
22-23 Jan 1927
31 Mar 1929
7-8 Jan 1930
1-2 Mar 1932
7-8 Jan 1937
31 Dec 1947–
1 Jan 1948
10 Jan 1949
8 Dec 1956
20-22 Jan 1959
26-27 Jan 1967
Radial
thickness
of ice
on wire
(inches)
Weight of
Ratio of
ice (oz)
ice weight on 1 ft of
to weight standard
of 0.25(#12)
inch twig
wire
City
State
section
Springfield
Decatur
La Salle
Marengo
Springfield
Cairo
Moline
Carlinville
Galena
Quincy
WSW
C
NE
NE
WSW
SE
NW
WSW
NW
W
72
Chicago
NE
40
Macomb
Alton
Urbana
Urbana
W
WSW
E
E
11
0.5
2.0
1.6
1.2
1.1
0.5
1.2
0.5
1.5
15:1
1.0
0.8
0.5
0.7
1.7
12:1
17:1
12
8
2
Figure 31. Greatest thickness of ice on wire in each square
during 1929-1937
37
nesses of ice exceeding 0.25 inch occurred in all parts of
Illinois during this 9-year period. They further reveal
that points in the south-central and northern portions of
the state had maximum glaze thicknesses exceeding 0.75
inch, although the area1 variations shown in figure 31
may largely be due to the short sampling period.
Until more precise data are collected, it is believed
that the point design values for extreme glazing are
equivalent anywhere in Illinois. The best estimates for
these extremes could be the highest values shown in table
28. Thus, wires anywhere in Illinois could experience
glaze with a radial thickness of 2 inches and a weight of
72 ounces per foot. However, Seeley 26 reported that the
glaze weight on wires from a I922 storm in Michigan
exceeded 176 ounces per foot, and that the ratio of ice
weight to twig weight was 32:1. Root 27 reported a ratio
of 15:1 at Springfield in 1924, and 17:1 was measured in
the center of the 1967 glaze storm (table 28). Tree damage from glaze increases with tree size, and the weight of
ice on a 50-foot high evergreen tree has been estimated
to be 50 tons. 7
Glaze-Wind
Relationships
Strong winds both while glaze is occurring and after a
storm greatly increase the amount of tree and wire damage. Bennett 7 has found that tree damage from glaze is
usually minor unless accompanied by strong winds. The
joint effect of ice and wind on wires greatly increases the
transverse loading (for wires normal to the wind), and
also upsets wire equilibrium, producing erratic movements and breakage.
Wind data during and after severe glaze storms are
often not collected because the glaze accumulation halts
the mechanical operation of the wind measuring equipment. This situation occurred in the two severest glaze
storms at Urbana (table 28).
The average hourly wind speeds measured during 542
hours when freezing rain or drizzle were occurring at
Urbana were calculated from all available data from the
1952-1967 period. These data showed that during 5 percent of the time (hours) wind speeds were under 5 mph;
in 19 percent of the time speeds were 6-10 mph; in 51
percent, 11-15 mph; in 21 percent, 16-20 mph; and in 4
percent 21-35 mph.
Winds after a severe glaze storm has ended, and when
accumulations are at a maximum, are equal or greater in
importance for damage production than winds during the
storm. In studying wind effects on glaze-loaded wires,
the Association of American Railroads 25 concluded that
maximum wind gusts were not as significant (harmful) a
measure of wind damage potential as were the speeds
sustained over 5-minute periods. In table 29, the maximum 5-minute wind speeds measured in the after-storm
periods of 148 glaze storms throughout the country during 1926-1937 are summarized for 5-mph intervals. 7 The
38
Table 29. Summary of Maximum 5-Minute Wind
Speeds Occurring after 148 Glaze Storms
Wind
speed
intervals
(mph)
Number of
cases
Number of cases when
radial thickness of ice
was 0.25-inch or more
0–4
5–9
10–14
15–19
20–24
25–29
30–34
35–39
40–44
45–49
50–54
1
17
35
46
27
10
6
2
1
2
1
0
2
3
15
6
3
1
1
0
1
0
Total
148
32
Table 30. Wind-Glaze Thickness Relations for Five
Periods of Greatest Speed and Greatest Thickness
Rank
1
2
3
4
5
Five periods when five
fastest 5-minute speeds
were registered
Speed
Ice thickness
(mph)
(inches)
50
46
45
40
35
0.19
0.79
0.26
0.30
0.78
Five periods when five
greatest ice thicknesses
were measured
Speed
Ice thickness
(inches)
(mph)
2.87
1.71
1.50
1.10
1.00
30
18
21
28
18
average of these 5-minute maximums was 17.5 mph, and
33 percent of the occurrences (cases) had speeds of 20
mph or greater. Shown with each interval in table 29 is
the number of times the glaze accumulation on wires had
a radial thickness of 0.25 inch or more. Twenty-seven of
the 32 cases with 0.25-inch or thicker glaze were with a
5-minute wind of 15 mph or faster. The average velocity
of the 32 cases was 20 mph.
Specific glaze thickness data for the five fastest 5minute speeds, and the speeds with the five greatest
measured glaze thicknesses, are shown in table 30.
Although these extremes were collected from various
locations throughout the United States in an 11-year
study period, they are considered applicable design
values for any location in Illinois. In a few unusual glaze
storms ( figure 20), the high wind speeds with glaze came
from associated thunderstorms.20 Although the amount
of wind data with and after glaze storms is small, the
available data suggest that moderate speeds are most
prevalent. Speeds of 25 mph or higher are not unusual,
and there have been 5-minute winds in excess of 40 mph
with glaze thicknesses of 0.5 inch or more.
Average Frequency of Days with
Heavy Snow, Glaze, and Sleet
A part of the design problem for severe winter storms
concerns the point frequencies of days with the various
conditions associated with these storms. Not all days with
glaze or sleet at a point are associated with severe storms,
a. Annual number of days with
1 inch or more of snowfall
b. Number of 2-day periods
with snowfall of 4 inches
or more in an average
lo-year period
d. Annual number of days
with glaze
c. Annual number of days
with 3 inches or more
of snow on the ground
e. Annual number of days
with sleet
Figure 32. Average patterns of days with snowfall, snow depth, glaze, and sleet, based on point data
39
Figure 33. Snowfall area-depth curves for average monthly snowstorms
40
but many are. Sixty-two Illinois stations 5 with weather
records for the 1901-1962 period were analyzed to determine point averages for the number of days with 1)
l-inch or heavier snowfalls, 2) 3-inch or deeper depths of
snow on the ground, 3) glaze, and 4) sleet. The number
of 2-day periods with 4 inches or more snowfall also was
calculated from these records, and the resulting average
patterns for all five conditions are shown in figure 32.
The distribution of heavy snowfall days and deep
snow-depth days (figures 32a, b, c) all exhibit latitudinal
distributions with points in northern Illinois experiencing
more than three times as many days, on the average, as
those in extreme southern Illinois. The highest point
values occur in the extreme northwestern corner of the
state. The glaze and sleet patterns (figures 32d and e)
both have a maximum oriented WSW-ENE across central Illinois, and their lowest values in extreme southern
Illinois.
AREAL DATA
Area-Depth Relations
A frequently used measure of the volume of precipitation over an area is the area-depth relationship. Curves
based on these relations allow determination of the area
covered by any volume of rain or snow.
Average monthly area-depth relations were developed
for all winter storms (269 total) that produced 6 inches
or more snowfall. The monthly area-depth curves in
figure 33 reveal differences that can be related to calculations of average snow removal and water input by these
severe storms. The average February storm covers considerably greater area with snow above 4 inches, whereas
the typical December storm is the most widespread for
depths of 1 to 4 inches. The average storm in April is
least extensive at most depths. Area1 differences in depths
less than 6 inches are considerable. The 6-inch depth in
the average storm in February covers 8500 square miles;
that in January is 8200 square miles; March, 8000; December, 7200; November, 6600; and that in April, only 5200
square miles. The shapes of the monthly average curves
reveal the tendency for heavy snowfalls (8 inches or
more) to be relatively small in area1 extent as compared
with the extent of the area experiencing between 1 and
8 inches.
The most extensive snowstorms in each month, defined
as those covering the greatest area of Illinois with 1 inch
or more snowfall, were used to construct the area-depth
curves in figure 34. The curve for the design storm on
27-28 February 1900 (figure 21) reveals very extensive
coverage at all depths above 5 inches. The extensive
storms in November, December, and January have areadepth relationships that are somewhat alike. The 11-12
January 1918 storm exceeded all other storms in area1
coverage in the 1- to 4-inch depths. Although the most
widespread storms in March and April produced 1 inch
or more snowfall over most of Illinois, these storms did
not produce exceptionally heavy snowfall over wide areas.
The storm snowfall data for each of the three damage
classes (minor, moderate, and extreme) were used to
develop average area-depth relationships for each class
(figure 35). The average curves show that at any depth,
Figure 34. Snowfall Oreo-depth curves for most widespread storms
in each month, 1900-1960
the area covered by a storm resulting in minor damage
(less than $10,000) was considerably less than that with
the more damaging storms. There was less difference
between the curve for the moderately damaging storm
and that of the extremely damaging storm, but at any
given depth the extreme storm had slightly more area of
snowfall. This suggests that the damage difference between these two classes was not a function of the total
snowfall. The degree of damage for a storm was more
likely a result of other factors, including the strength of
the wind (drifting), the severity of the associated glaze,
and the location of storm center (largely in rural or in
urban areas).
Average area-depth relations were developed for the
6-inch or greater snowstorms in each of the five major
41
Figure 35. Average snowfall area-depth curves for different
classes of damage
Figure 36. Average snowfall area-depth curves for different
weather types
weather types. The average curves (figure 36) reveal
that the average Alberta-derived storm (type 1) produces
less extensive snowfall at most depths. The type 2 storm,
which is the most prevalent type (table 16), has on the
average the most extensive areas with snowfall exceeding
5 inches. At lower depths, 1 to 3 inches, the average
type 4 storm has greater area1 coverage.
The greatest difference in the area-depth curves of the
five weather types is found in the values below 6 inches.
There is little significant difference (less than 1000 square
miles) in their area-depth values above 10 inches. The
curves at the l-inch depth show that, on the average, the
type 4 and 5 conditions produce the most widespread
snowstorms in Illinois (35,000 to 37,500 square miles),
whereas the type 1 and type 3 produce the least extensive storms (27,000 to 28,500 square miles). Interestingly, the most widespread storms (types 4 and 5) were
quite infrequently related to the more damaging storms
(table 18). These results further illustrate the relatively
inexact relationship between the area1 extent of snowfall
and degree of storm damage.
The glaze storm data on area1 extent (table 22) and
orientation were also used in developing the storm model
(figure 37). The glaze-snowstorms were analyzed to determine the average distance between the major axis of
the glaze area and the southern boundary of the 6-inch
snowfall area. As shown, the average distance was 50
miles. The length-width ratio of the glaze areas was calculated to be 4:1 and this was used with the average
area1 extent of glaze (8500 square miles) to derive an
area 184 miles long by 46 miles wide.
Frequencies of stations reporting sleet in each winter
storm were used to determine the average area1 extent
of sleet with the severe winter storms. The average extent was 11,800 square miles in a rectangular area, with
a length 2.4 times its width. The sleet area usually was
oriented with the glaze area, with its southern edge
through the central portion of the glaze area (figure 37).
The pattern shown in figure 37 is considered to model
the sizes and area1 relationships of the basic features of
a severe winter storm in Illinois. The synoptic weather
condition most likely to produce this storm is a low developing in Colorado that initially moves southeastward
but then northeastward up the Ohio River Valley. However, only 10 percent of the lows that develop in Colorado during the cold season produce a severe storm in
Illinois. The southern edge of the 4-inch snowfall area
(figure 37) is usually 50 to 75 miles north of the attendant surface low. 28
Storm Model
Several average features of the 304 severe winter
storms were utilized to develop a dimensional model of
the storms. The average values of snowfall areas (table
6) from the 269 storms with 6 inches or more were used
along with the 3:1 ratio of the median values for length
and width of storm cores to develop the snowfall dimensions in the model (figure 37). The preferred orientation
of all storm cores, 250 degrees, was used to orient the
model, and other statistics relating to storm motion, duration, and initiation are shown.
42
Regional Probabilities of Severe Storm Centers
The frequency of 6-inch snowfall cores expected during
an average 10-year period is shown in figure 38a for the
nine districts of Illinois. Both northern districts can ex-
Figure 37. Model of severe winter storm in Illinois
pect two to three such storm cores in a year, whereas
each of the two southernmost districts can expect only
one a year.
The frequency of damaging glaze areas in three state
regions (figure 38b) during an average 10-year period
reveals a greater likelihood of their occurrence in central
Illinois. A 23-year study of ice on telephone wires 7 provided data for portrayal of area1 frequencies of storms
producing thick (0.75-inch) glaze on wires (figure 38c).
This pattern shows that most of Illinois has an equal
likelihood for one glaze storm of this severity at least
once every three years. Only in extreme southern Illinois
is the heavy glaze storm likelihood considerably less.
Regional Probabilities for Successive Storms
The chance for a 6-inch snowfall core to occur in an
area after one has just occurred was determined for varying periods of time up to 30 days. Probabilities developed
for each of the nine state districts (table 31) show considerable area1 variations. For instance, probabilities for
a second storm within 20 days in the WSW and SE districts are nearly 30 percent, whereas the probability in
the W district is only 12 percent. For all time periods,
chances for a following storm (core) are greatest in the
WSW, SW, SE, and NE districts. Chances for successive
storm cores are poorest in the W and C districts.
Persistence of Severe Snow and Glaze
Conditions in Storm Centers
The length of time (days) after a storm ended until
the heavy snow in the storm cores began melting and was
Table 31. Probability That a 6-Inch Storm Core
Will Occur in a District after One Has Ended
Time of
occurrence
of second storm
after end of
first storm
Within
Within
Within
Within
Within
Within
5
10
15
20
25
30
days
days
days
days
days
days
Probability (percent) in each district
NW
NE
W
C
E
6
10
15
17
19
20
9
15
19
22
25
27
1
7
10
12
13
14
6
13
15
16
17
18
6
11
14
18
20
22
WSW ESE
13
19
24
28
29
30
7
14
15
17
18
18
SW
SE
9
17
18
20
25
26
14
22
26
28
32
32
43
a. Number of 6-inch snowfall
cores in an average
10-year period
b. Number of damaging glaze
areas expected in an
average 10-year period
c. Area1 frequencies of
storms with 0.75-inch
or thicker ice on wires
Figure 38. Regional frequencies of severe storm centers
not a local problem could be determined for 192 of the
269 storms with 6-inch or greater snowfalls. The distribution of lengths (table 32) reveals that snowfall persistence in all months except January was usually less
than 3 days. Persistence seldom exceeded 15 days for
storms in most months, and the longest value was 32 days.
The average values (table 32) vary from less than 1 day
(April) to 5 days for January storms.
A similar analysis for the 92 glaze storms was based
on the persistence of glaze in the center (worst) of the
Table 32. Persistence of Deep Snow in Storm Cores
after Storm Ends
Persistence
(days)
1-3
A-7
8-15
≥ 16
Average
length (days)
Maximum
length (days)
44
Number of storms
Jan
Feb
Nov
Dec
Mar
Apr
12
1
1
0
23
10
8
1
21
13
9
8
28
7
6
6
29
1
2
2
4
0
0
0
2
3
5
4
2
<1
9
20
32
28
20
1
glaze area. In 66 storms (72 percent), the heaviest glaze
layers disappeared within 2 days; in 11 storms, in 3 to 5
days; in 8 storms, in 6 to 8 days; in 4 storms, in 9 to 11
days; and in 3 storms, from 12 to 15 days. Fifteen days
was the maximum persistence of glaze in any storm center. The average glaze persistence in each month was
less than 2 days.
Average persistence values were calculated for the
center (worst) part of the minor damage storms, the
center of the moderately damaging storms, and the center of the extremely damaging storms. The average value
for the minor storms was 5 days, that for the moderately
damaging class ($10,000-$200,000) was 8 days, and that
for the extremely damaging class was 10 days.
Persistence values for the five weather types also were
analyzed. The average values for type 1 (Alberta), type
4 (Colorado-east track), and type 5 (Texas-West Gulf)
were 2 to 3 days. The average value for type 2 (Colorado-south track) was 6 days, and that for type 3 (Colorado-north track) was 5 days. This indicates that, on
the average, storms of types 2 and 3 are followed by
colder air than are storms of the other types.
REFERENCES
1 Changnon, S. A., and G. E. Stout. 1957. Tornadoes
in Illinois. Illinois State Water Survey Circular 63.
2 Changnon, S. A. 1964. Climatology of damaging
lightning in Illinois. Monthly Weather Review v.
62: 115-119.
3 Brown, P. S. 1967. Crop-hail insurance. Society
of Chartered Property and Casualty Underwriters
Annals, p. 233-277.
4 U. S. Department of Commerce. 1900-1960. Illinois climatological data. Weather Bureau v. 6-65.
5 Changnon, S. A. 1957. Illinois cooperative project
in climatology, third progress report. Illinois State
Water Survey Circular 64.
6 Austin, J. M., and S. L. Hensel. 1956. Analysis of
freezing precipitation along the eastern North American coastline. Massachusetts Institute of Technology Lincoln Laboratory Technical Report 112.
7 Bennett, I. 1959. Glaze, its meteorology and climatology, geographical distribution, and economic
effects. U. S. Army Quartermaster Research and
Engineering Center Technical Report EP-105.
8 Illinois Cooperative Crop Reporting Service. 1958.
Illinois agricultural statistics. Bulletin 58-2.
9 Changnon, S. A. 1968. Climatology of hourly occurrences of selected atmospheric phenomena in
Illinois. Illinois State Water Survey Circular 93.
10 Klein, W. H. 1957. Principal tracks and mean
frequencies of cyclones and anticyclones in the
northern hemisphere. U. S. Weather Bureau Research Paper 40.
11 Petterssen, S. 1955. Cyclone development. University of Chicago Final Report Contract AF 19(604) -390.
12 McClain, E. P. 1960. Some effects of the Western
Cordillera of North America on cyclonic activity.
Journal of Meteorology v. 17:104-115.
13 Schallert, W. L. 1962. An investigation of Colorado cyclones. University of Chicago Contract
AF 19 (604) - 7230 Scientific Report 7.
16 Bowie, E. H., and R. H. Weightman. 1914. Types
of storms of the United States and their average
movement. Monthly Weather Review Supplement.
17 Ward, R. A. 1967. The great Midwest snowstorm
of January 1967 in the Chicago-Calumet area.
Weatherwise v. 20:67-70.
18 Williams, G. W. 1963. An occurrence of lake
snow, one of the direct effects of Lake Michigan
on the climate of the Chicago area. Monthly
Weather Review v. 91:465-467.
19 Changnon, S. A. 1968. Precipitation climatology
of Lake Michigan basin. Illinois State Water Survey
Bulletin 52.
20 Changnon, S. A. 1964. Freak winter snowstorms
in Illinois. Weatherwise v. 17: 272-276.
21 Hanks, H. H., D. R. De Harpporte, and E. C. Grueber. 1967. Snow forecasting procedures in the
central United States. Atmospheric Research and
Development Corporation Contract CWB 11353
Report.
22 Boggess, W. R., and F. W. McMillan. 1954. Cold
weather and glaze damage to forest plantations in
southern Illinois. University of Illinois Agricultural
Experiment Station Bulletin 574.
23 Riehl, H. 1952. Forecasting in the middle latitudes. Meteorological Monographs v. l (5) :47-51.
24 Potter, J. G. 1965. Water content of freshly fallen
snow. Canadian Department Transport, Meteorological Branch Circular 4232.
25 Association of American Railroads. 1955. Glaze
storm loading summary, 1927-28 to 1936-37.
26 Seeley, D. A. 1922. The great glaze storm of
February 21-23, 1922, in Michigan. Monthly
Weather Review v. 50:80-82.
27 Root, C. J. 1924. The glaze storm of December
17-18, 1924, in Illinois. Monthly Weather Review
v. 52 : 584-585.
14 Saucier, W. 1949. Texas-West Gulf cyclones.
Monthly Weather Review v. 77:219-231.
28 Hanks, H. H. 1966. Heavy snow forecasting in
the central United States. Atmospheric Research
and Development Corporation Contract CWB 11098
Final Report.
15 Namias, J., and P. E. Clapp. 1946. Normal fields
of convergence and divergence at the 10,000-foot
level. Journal of Meteorology v. 3:14-22.
29 Bryson, R. A. 1966. Air masses, streamlines and
the boreal forest. Geographical Bulletin v. 8:228269.
45